412 editable keywords across 74 works. The pool includes broad themes, specific mechanisms, datasets, representations, and useful search aliases. It deliberately exceeds the plotted taxonomy.
Edit _data/research_map_keywords.yml to rename keywords or add aliases. Edit each work’s keywords in _data/research_map.yml to change assignments. Keywords and aliases are retained for editing and analysis; editing memberships changes the map’s categories. No connection is inferred automatically from this list.
Editing guide · Complete analysis · Homepage map
Keyword index
2D–3D views
ID: views_2d3d.
accidental correct answers
ID: accidental_correct_answers.
action subsampling
ID: action_subsampling.
adaptive grid
ID: adaptive_grid.
Adaptive resolution
ID: adaptive_resolution · Map category: c_adaptive_resolution.
DNAChunker, Spherical neural fields, Adaptive-grid exploration, K²-tree graph generation, Wavelet graph diffusion, Layer-adaptive pruning.
adaptive stages
ID: adaptive_stages.
Adaptive trajectories
ID: adaptive_trajectories · Map category: c_adaptive_trajectories.
MELD, Pessimistic backward policy, EPIC, Learning what to defer.
ADMET
ID: admet.
Search aliases: absorption distribution metabolism excretion toxicity.
adsorbate
ID: adsorbate.
Adsorption & host–guest modeling
ID: adsorption_hostguest_modeling · Map category: c_adsorption.
affinity proxy
ID: affinity_proxy.
Antibody sequence–structure decoupling.
agent profiles
ID: agent_profiles.
agent safety
ID: agent_safety.
AGG-WL
ID: agg_wl.
alanine dipeptide
ID: alanine_dipeptide.
Algorithm discovery
ID: algorithm_discovery · Map category: c_algorithm_discovery.
All-atom modeling
ID: all_atom_modeling · Map category: c_all_atom.
AtomMOF, TriProRep, Packora, ReBind.
Amortized inference / generation
ID: amortized_inference_generation · Map category: c_amortization.
AtomMOF, MADField, Latent veracity inference, Energy-based generator matching, Search-guided diffusion samplers, MOFFlow, TPS-DPS, Adaptive Teachers, Iterated energy-based flow matching, Pessimistic backward policy, LED-GFN, BootGen.
antibody
ID: antibody.
Antibody sequence–structure decoupling.
any-to-any generation
ID: any_to_any_generation.
apprentice
ID: apprentice.
Genetic expert-guided learning.
ARC-MOF
ID: arc_mof.
attributed grammar
ID: attributed_grammar.
Auxiliary teachers
ID: auxiliary_teachers · Map category: c_auxiliary_teacher.
Adaptive Teachers, Learning with a biased committee, Learning from Failure.
available reactants
ID: available_reactants.
average velocity
ID: average_velocity.
backtracking
ID: backtracking.
backward policy
ID: backward_policy.
bandwidth
ID: bandwidth.
batch effects
ID: batch_effects.
Benchmarking
ID: benchmarking · Map category: c_benchmarking.
Co-folding representations, TriProRep, scTrilemma, Packora, MaskGXT / HACO, VibeProteinBench, AdaPert, PBio-Agent / LincsQA, QHFlow2, BioEmu-CV, RL4CO, Multi-bias robust learning, Concept intervention analysis.
Benchmarks & evaluation
ID: benchmarks_evaluation · Map category: m_benchmark.
Co-folding representations, TriProRep, Packora, VibeProteinBench, PBio-Agent / LincsQA, QHFlow2, RL4CO, Concept intervention analysis.
bias attributes
ID: bias_attributes.
bias-conflicting samples
ID: bias_conflicting_samples.
Learning with a biased committee.
biased committee
ID: biased_committee.
Learning with a biased committee.
biased learner
ID: biased_learner.
bioactivity
ID: bioactivity.
BioEmu
ID: bioemu.
blind test
ID: blind_test.
blossom contraction
ID: blossom_contraction.
blossom expansion
ID: blossom_expansion.
Boltz2
ID: boltz2.
bootstrapping
ID: bootstrapping.
Energy-based generator matching.
BP correction
ID: bp_correction.
breadth-first exploration
ID: breadth_first_exploration.
building-block library
ID: building_block_library.
candidate inventory
ID: candidate_inventory.
carbon capture
ID: carbon_capture.
cardinality
ID: cardinality.
Unsupervised combinatorial optimization.
catalyst
ID: catalyst.
causal influence diagram
ID: causal_influence_diagram.
Causal reasoning & interventions
ID: causal_reasoning_interventions · Map category: c_causal_reasoning.
Causal Influence Prompting, Concept intervention analysis.
cDFT
ID: cdft.
Search aliases: classical density functional theory.
CDR-H3
ID: cdr_h3.
Antibody sequence–structure decoupling.
cell-line transfer
ID: cell_line_transfer.
Cells & perturbations
ID: cells_perturbations · Map category: d_cells.
scTrilemma, AdaPert, PBio-Agent / LincsQA.
chemist agents
ID: chemist_agents.
CNN-to-MLP
ID: cnn_to_mlp.
Variational information distillation.
collective variables
ID: collective_variables.
Combinatorial optimization
ID: combinatorial_optimization · Map category: c_combinatorial.
RL4CO, RxnFlow, Symmetric replay training, Unsupervised combinatorial optimization, Learning what to defer, Odd-cycle matching BP, Blossom belief propagation.
compact tree operation
ID: compact_tree_operation.
Spanning-tree molecular generation.
Competing objectives
ID: competing_objectives · Map category: c_tradeoffs.
scTrilemma, Antibody sequence–structure decoupling, Multi-bias robust learning.
composition objective
ID: composition_objective.
Antibody sequence–structure decoupling.
composition ordering
ID: composition_ordering.
Compression
ID: compression · Map category: c_compression.
DNAChunker, Spherical neural fields, Gap-encoded edge lists, K²-tree graph generation, Layer-adaptive pruning, Variational information distillation, Bucket renormalization.
Compression & compact coding
ID: compression_compact_coding · Map category: m_compression.
Gap-encoded edge lists, K²-tree graph generation, Layer-adaptive pruning, Variational information distillation.
concept bottleneck
ID: concept_bottleneck.
Concept intervention analysis.
Conditional generation
ID: conditional_generation · Map category: c_conditioning.
Packora, Multimodal Crystal Flow, CatFlow, BioEmu-CV, MOFFlow-2, QHFlow, RxnFlow, MOFFlow, Structurally diverse molecular LLMs, Structured node diffusion.
conformation
ID: conformation.
conformer
ID: conformer.
Search-guided diffusion samplers.
constrained sampling
ID: constrained_sampling.
Constraint-aware design
ID: constraint_aware_design · Map category: c_constraints.
MELD, MOFFlow-2, RxnFlow, Antibody sequence–structure decoupling, Unsupervised combinatorial optimization, Adaptive-grid exploration, EPIC, Structured node diffusion, Difficult, not too different, Spanning-tree molecular generation, RoMA, Self-improved retrosynthesis, RetCL, Learning what to defer, Odd-cycle matching BP, Blossom belief propagation.
Context selection
ID: context_selection · Map category: c_context_selection.
AdaPert, PBio-Agent / LincsQA, Causal Influence Prompting.
context-conditioned prior
ID: context_conditioned_prior.
context-sensitive cost
ID: context_sensitive_cost.
EPIC.
continuous sampling
ID: continuous_sampling.
Contrastive learning
ID: contrastive_learning · Map category: c_contrastive.
Control & multi-agent learning
ID: control_multi_agent_learning · Map category: d_deep.
Adaptive-grid exploration, Planner-guided imitation, Disentangled risk-sensitive MARL, RoMA.
control variates
ID: control_variates.
corruption schedule
ID: corruption_schedule.
MELD.
cosmic microwave background
ID: cosmic_microwave_background.
Credit assignment
ID: credit_assignment · Map category: c_credit_assignment.
Search aliases: reward decomposition.
Co-folding representations, LED-GFN.
CRISPR
ID: crispr.
Cross-modal learning
ID: cross_modal_learning · Map category: c_cross_modal.
Co-folding representations, TriProRep, VibeProteinBench, Multimodal Crystal Flow, HoliMol.
crossover
ID: crossover.
Genetic expert-guided learning.
crystal structure prediction
ID: crystal_structure_prediction.
CUB
ID: cub.
Concept intervention analysis.
Curriculum & training difficulty
ID: curriculum_training_difficulty · Map category: c_curriculum.
PBio-Agent / LincsQA, CleanMol, Adaptive Teachers, Learning with a biased committee, Difficult, not too different, Learning from Failure.
cutting plane
ID: cutting_plane.
Data augmentation
ID: data_augmentation · Map category: m_augmentation.
EPIC, Difficult, not too different, Self-improved retrosynthesis.
decision utilities
ID: decision_utilities.
deferred decisions
ID: deferred_decisions.
Density / field representations
ID: density_field_representations · Map category: c_density_fields.
MADField, Spherical neural fields, Gaussian plane-wave neural operator, Wavelet graph diffusion.
Derandomization
ID: derandomization · Map category: c_derandomization.
Unsupervised combinatorial optimization.
description-guided generation
ID: description_guided_generation.
Structurally diverse molecular LLMs.
Design-bench
ID: design_bench.
RoMA.
differential expression
ID: differential_expression.
diffusion Learner
ID: diffusion_learner.
Search-guided diffusion samplers.
Diffusion models
ID: diffusion_models · Map category: m_diffusion.
MELD, BioEmu-CV, Energy-based generator matching, Search-guided diffusion samplers, TPS-DPS, Wavelet graph diffusion, Structured node diffusion.
Diffusion Transformer
ID: diffusion_transformer.
Discrete–continuous states
ID: discretecontinuous_states · Map category: c_mixed_state.
Energy-based generator matching.
distributed algorithm
ID: distributed_algorithm.
distributional value
ID: distributional_value.
Disentangled risk-sensitive MARL.
Diversity & mode coverage
ID: diversity_mode_coverage · Map category: c_diversity.
MaskGXT / HACO, INDIBATOR, Search-guided diffusion samplers, RxnFlow, Adaptive Teachers, Antibody sequence–structure decoupling, Structurally diverse molecular LLMs, Pessimistic backward policy, Local Search GFlowNets, BootGen.
docking
ID: docking.
double well
ID: double_well.
Iterated energy-based flow matching.
edge list
ID: edge_list.
electron density
ID: electron_density.
Gaussian plane-wave neural operator.
electronic Hamiltonian
ID: electronic_hamiltonian.
Energy / reward guidance
ID: energy_reward_guidance · Map category: c_energy_guidance.
energy and forces
ID: energy_and_forces.
Energy-based sampling
ID: energy_based_sampling · Map category: m_energy.
Energy-based generator matching, Search-guided diffusion samplers, TPS-DPS, Adaptive Teachers, Iterated energy-based flow matching, MCMC + belief propagation.
enhanced sampling
ID: enhanced_sampling.
environment uncertainty
ID: environment_uncertainty.
Disentangled risk-sensitive MARL.
equivariant bias
ID: equivariant_bias.
Equivariant models
ID: equivariant_models · Map category: m_equivariant.
AtomMOF, Packora, Riemannian MeanFlow, QHFlow2, CatFlow, MOFFlow-2, QHFlow, MOFFlow, TPS-DPS, ReBind, Antibody sequence–structure decoupling, Gaussian plane-wave neural operator.
exchange-correlation
ID: exchange_correlation.
Gaussian plane-wave neural operator.
expert rationale
ID: expert_rationale.
Exploration–exploitation
ID: explorationexploitation · Map category: c_exploration.
Search aliases: exploration-exploitation; exploration exploitation.
Search-guided diffusion samplers, Adaptive Teachers, Pessimistic backward policy, Symmetric replay training, Adaptive-grid exploration, Local Search GFlowNets, Disentangled risk-sensitive MARL, Genetic expert-guided learning.
expressivity
ID: expressivity.
facility location
ID: facility_location.
Unsupervised combinatorial optimization.
Factorization & decomposition
ID: factorization_decomposition · Map category: c_factorization.
CatFlow, MOFFlow-2, MOFFlow, Antibody sequence–structure decoupling, Gaussian plane-wave neural operator, HoliMol, LED-GFN, Disentangled risk-sensitive MARL, Bucket renormalization, Gauged variational inference, Gauged mini-bucket elimination.
failure weighting
ID: failure_weighting.
fairness
ID: fairness.
Concept intervention analysis.
Few-step generation
ID: few_step_generation · Map category: c_few_step.
Feynman–Kac steering
ID: feynmankac_steering.
flexible building block
ID: flexible_building_block.
flow map
ID: flow_map.
Flow matching
ID: flow_matching · Map category: m_flow.
Search aliases: continuous flow.
AtomMOF, Packora, Multimodal Crystal Flow, Riemannian MeanFlow, CatFlow, Energy-based generator matching, MOFFlow-2, QHFlow, MOFFlow, Iterated energy-based flow matching.
Forney graph
ID: forney_graph.
forward reaction augmentation
ID: forward_reaction_augmentation.
forward reconstruction
ID: forward_reconstruction.
fragment contrast
ID: fragment_contrast.
free energy
ID: free_energy.
frequency decomposition
ID: frequency_decomposition.
Gaussian plane-wave neural operator.
full-atom token
ID: full_atom_token.
gap encoding
ID: gap_encoding.
gas uptake
ID: gas_uptake.
gauge optimization
ID: gauge_optimization.
Gauged mini-bucket elimination.
gauge transformation
ID: gauge_transformation.
Gaussian mixture
ID: gaussian_mixture.
Iterated energy-based flow matching.
Gaussian orbital
ID: gaussian_orbital.
Gaussian plane-wave neural operator.
GCMC
ID: gcmc.
Search aliases: grand canonical Monte Carlo.
GEEL
ID: geel.
GEGL
ID: gegl.
Genetic expert-guided learning.
gene regulation
ID: gene_regulation.
General learning & inference
ID: general_learning_inference.
Riemannian MeanFlow, Latent veracity inference, Energy-based generator matching, Search-guided diffusion samplers, Adaptive Teachers, Iterated energy-based flow matching, Pessimistic backward policy, Multi-bias robust learning, LED-GFN, Local Search GFlowNets, Concept intervention analysis, Learning with a biased committee, RoMA, Layer-adaptive pruning, Learning from Failure, Variational information distillation, Bucket renormalization, Gauged variational inference, Gauged mini-bucket elimination, Odd-cycle matching BP, MCMC + belief propagation, Blossom belief propagation.
generator matching
ID: generator_matching.
Energy-based generator matching.
Genetic algorithms
ID: genetic_algorithms · Map category: c_genetic_algorithm.
Search aliases: genetic algorithm; evolutionary search.
Genetic expert-guided learning.
genomic segmentation
ID: genomic_segmentation.
GFlowNets
ID: gflownets · Map category: m_gfn.
Search aliases: GFN; generative flow network.
RxnFlow, Adaptive Teachers, Pessimistic backward policy, LED-GFN, Local Search GFlowNets.
global graph
ID: global_graph.
GLUE
ID: glue.
goal-conditioned policy
ID: goal_conditioned_policy.
goal-conditioned RL
ID: goal_conditioned_rl.
graph edit path
ID: graph_edit_path.
EPIC.
Graph neural networks
ID: graph_neural_networks · Map category: m_graph.
Search aliases: GNN.
AdaPert, QHFlow2, MELD, ReBind, Non-backtracking GNNs, HoliMol, EPIC, Wavelet graph diffusion, Structured node diffusion, RetCL, Learning what to defer.
graph planner
ID: graph_planner.
graph rewiring
ID: graph_rewiring.
Graph structure & transformations
ID: graph_structure_transformations · Map category: c_graph_structure.
AdaPert, CleanMol, Molecular Structural Reasoning, ReBind, Non-backtracking GNNs, Adaptive-grid exploration, EPIC, Gap-encoded edge lists, K²-tree graph generation, Wavelet graph diffusion, Spanning-tree molecular generation, Odd-cycle matching BP, Blossom belief propagation.
graph wavelet
ID: graph_wavelet.
Graphs & discrete problems
ID: graphs_discrete_problems.
RL4CO, Non-backtracking GNNs, Symmetric replay training, Unsupervised combinatorial optimization, EPIC, LED-GFN, Gap-encoded edge lists, K²-tree graph generation, Wavelet graph diffusion, Structured node diffusion, Spanning-tree molecular generation, Learning what to defer, Odd-cycle matching BP, Blossom belief propagation.
group loss
ID: group_loss.
HACO
ID: haco.
half-integrality
ID: half_integrality.
hard negatives
ID: hard_negatives.
hard-core model
ID: hard_core_model.
hardware design
ID: hardware_design.
heterogeneous architecture
ID: heterogeneous_architecture.
Variational information distillation.
hierarchical tokens
ID: hierarchical_tokens.
high-loss regions
ID: high_loss_regions.
HNeR
ID: hner.
homodimer
ID: homodimer.
Importance sampling
ID: importance_sampling · Map category: c_importance_sampling.
Energy-based generator matching, Iterated energy-based flow matching.
in silico validation
ID: in_silico_validation.
Information routing
ID: information_routing · Map category: c_information_routing.
scTrilemma, Variational information distillation.
interatomic potential
ID: interatomic_potential.
intermediate reasoning
ID: intermediate_reasoning.
Molecular Structural Reasoning.
interpolation
ID: interpolation.
EPIC.
intervention budget
ID: intervention_budget.
Concept intervention analysis.
Invariance & spurious factors
ID: invariance_spurious_factors · Map category: c_invariance.
scTrilemma, Multi-bias robust learning, Learning with a biased committee, Learning from Failure.
Ising
ID: ising.
jump process
ID: jump_process.
Energy-based generator matching.
Knowledge distillation
ID: knowledge_distillation · Map category: c_distillation.
Co-folding representations, BootGen, Planner-guided imitation, Variational information distillation.
knowledge exchange
ID: knowledge_exchange.
Learning with a biased committee.
knowledge graph
ID: knowledge_graph.
K²-tree
ID: k2_tree.
label corruption
ID: label_corruption.
EPIC.
Language & scientific reasoning
ID: language_scientific_reasoning.
MaskGXT / HACO, VibeProteinBench, INDIBATOR, PBio-Agent / LincsQA, Latent veracity inference, CORE-PO, MT-Mol, CleanMol, Causal Influence Prompting, Molecular Structural Reasoning, Structurally diverse molecular LLMs, Difficult, not too different.
Language / sequence models
ID: language_sequence_models · Map category: m_language.
TriProRep, MaskGXT / HACO, VibeProteinBench, INDIBATOR, PBio-Agent / LincsQA, Latent veracity inference, DNAChunker, MOFFlow-2, CORE-PO, MT-Mol, CleanMol, Causal Influence Prompting, Molecular Structural Reasoning, Structurally diverse molecular LLMs, Gap-encoded edge lists, K²-tree graph generation, Spanning-tree molecular generation.
latent bottleneck
ID: latent_bottleneck.
lattice
ID: lattice.
layerwise sparsity
ID: layerwise_sparsity.
lead optimization
ID: lead_optimization.
Learned / classical hybrids
ID: learned_classical_hybrids · Map category: c_hybrid_solver.
MADField, BioEmu-CV, QHFlow, Bucket renormalization, MCMC + belief propagation.
learned augmentation
ID: learned_augmentation.
learning difficulty
ID: learning_difficulty.
Lennard-Jones
ID: lennard_jones.
LincsQA
ID: lincsqa.
linear program
ID: linear_program.
Local feedback
ID: local_feedback · Map category: c_local_feedback.
Latent veracity inference, CORE-PO, MT-Mol, Molecular Structural Reasoning, LED-GFN, Concept intervention analysis.
Local search
ID: local_search · Map category: c_local_search.
Local Search GFlowNets, RoMA, Genetic expert-guided learning, Learning what to defer.
local smoothness
ID: local_smoothness.
RoMA.
locally decomposable objective
ID: locally_decomposable_objective.
log-variance loss
ID: log_variance_loss.
Long-range propagation
ID: long_range_propagation · Map category: c_long_range.
loop calculus
ID: loop_calculus.
low confidence
ID: low_confidence.
low-budget optimization
ID: low_budget_optimization.
low-degree atoms
ID: low_degree_atoms.
low-rank projection
ID: low_rank_projection.
lower bound
ID: lower_bound.
Gauged mini-bucket elimination.
LP relaxation
ID: lp_relaxation.
magnitude pruning
ID: magnitude_pruning.
Manifold geometry
ID: manifold_geometry · Map category: c_manifolds.
Riemannian MeanFlow, MOFFlow-2, MOFFlow, Spherical neural fields.
masked diffusion
ID: masked_diffusion.
MELD.
masked generative modeling
ID: masked_generative_modeling.
masked language model
ID: masked_language_model.
MaskGXT
ID: maskgxt.
matched sample budgets
ID: matched_sample_budgets.
Materials & electronic structure
ID: materials_electronic_structure.
AtomMOF, Packora, MaskGXT / HACO, MADField, Multimodal Crystal Flow, QHFlow2, CatFlow, MOFFlow-2, QHFlow, MOFFlow, Gaussian plane-wave neural operator.
maximum coverage
ID: maximum_coverage.
Unsupervised combinatorial optimization.
maximum independent set
ID: maximum_independent_set.
LED-GFN, Learning what to defer.
maximum-weight matching
ID: maximum_weight_matching.
MCMC
ID: mcmc · Map category: c_mcmc.
Search aliases: Markov chain Monte Carlo.
Search-guided diffusion samplers, MCMC + belief propagation.
MCMC Searcher
ID: mcmc_searcher.
Search-guided diffusion samplers.
mean field
ID: mean_field.
mechanism of action
ID: mechanism_of_action.
Message passing
ID: message_passing · Map category: c_message_passing.
Non-backtracking GNNs, Structured node diffusion, Odd-cycle matching BP, MCMC + belief propagation, Blossom belief propagation.
metal library
ID: metal_library.
Method transfer
ID: method_transfer · Map category: c_transfer.
mini-bucket
ID: mini_bucket.
minimax Pareto
ID: minimax_pareto.
mode coverage
ID: mode_coverage.
model aggregation
ID: model_aggregation.
model distortion
ID: model_distortion.
Modular frameworks
ID: modular_frameworks · Map category: c_modularity.
MOF
ID: mof.
Search aliases: MOFs; metal-organic framework.
Mol-Instructions
ID: mol_instructions.
molecular crystal
ID: molecular_crystal.
molecular fragmentation
ID: molecular_fragmentation.
molecular graph reconstruction
ID: molecular_graph_reconstruction.
Molecular Structural Reasoning.
molecular validity
ID: molecular_validity.
MELD.
Molecules & drug discovery
ID: molecules_drug_discovery · Map category: d_molecules.
AtomMOF, Co-folding representations, Packora, INDIBATOR, QHFlow2, CatFlow, MELD, Search-guided diffusion samplers, MOFFlow-2, QHFlow, MT-Mol, CleanMol, Molecular Structural Reasoning, RxnFlow, TPS-DPS, ReBind, Adaptive Teachers, Structurally diverse molecular LLMs, Pessimistic backward policy, Gaussian plane-wave neural operator, Symmetric replay training, HoliMol, LED-GFN, Gap-encoded edge lists, K²-tree graph generation, Local Search GFlowNets, Spanning-tree molecular generation, Self-improved retrosynthesis, RetCL, Genetic expert-guided learning.
Multi-agent deliberation
ID: multi_agent_deliberation · Map category: c_deliberation.
INDIBATOR, PBio-Agent / LincsQA, MT-Mol.
Multi-fidelity learning
ID: multi_fidelity_learning · Map category: c_multifidelity.
MultiCelebA
ID: multiceleba.
mutation
ID: mutation.
Genetic expert-guided learning.
mutation resilience
ID: mutation_resilience.
mutual information
ID: mutual_information.
Variational information distillation.
natural-language design
ID: natural_language_design.
Neural fields & operators
ID: neural_fields_operators · Map category: m_fields.
MADField, Spherical neural fields, Gaussian plane-wave neural operator.
node sensitivity
ID: node_sensitivity.
node–edge coupling
ID: nodeedge_coupling.
non-backtracking
ID: non_backtracking.
non-bonded interactions
ID: non_bonded_interactions.
novelty reward
ID: novelty_reward.
Search-guided diffusion samplers.
OC20
ID: oc20.
odd-cycle constraint
ID: odd_cycle_constraint.
off-policy flow matching
ID: off_policy_flow_matching.
Iterated energy-based flow matching.
offline design
ID: offline_design.
offline model-based optimization
ID: offline_model_based_optimization.
RoMA.
oracle budget
ID: oracle_budget.
orbital alignment
ID: orbital_alignment.
orientation
ID: orientation.
oversquashing
ID: oversquashing.
partially observed labels
ID: partially_observed_labels.
partition function
ID: partition_function.
Iterated energy-based flow matching, Bucket renormalization, Gauged variational inference, MCMC + belief propagation.
penalized logP
ID: penalized_logp.
Genetic expert-guided learning.
perfect matching
ID: perfect_matching.
perturbation response
ID: perturbation_response.
PerturbQA
ID: perturbqa.
Physical observables
ID: physical_observables · Map category: c_physical_observables.
QHFlow2, QHFlow, Gaussian plane-wave neural operator.
Physics-informed priors
ID: physics_informed_priors · Map category: c_physics_priors.
CatFlow, ReBind, Gaussian plane-wave neural operator.
plane wave
ID: plane_wave.
Gaussian plane-wave neural operator.
planner-free execution
ID: planner_free_execution.
PMO-1K
ID: pmo_1k.
pocket conditioning
ID: pocket_conditioning.
policy optimization
ID: policy_optimization.
polymer
ID: polymer.
MELD.
polymorph
ID: polymorph.
polymorph coverage
ID: polymorph_coverage.
potential function
ID: potential_function.
primacy bias
ID: primacy_bias.
Search-guided diffusion samplers.
primitive cell
ID: primitive_cell.
priority queue
ID: priority_queue.
Genetic expert-guided learning.
Probabilistic inference
ID: probabilistic_inference · Map category: m_inference.
scTrilemma, Latent veracity inference, Unsupervised combinatorial optimization, LED-GFN, Structured node diffusion, Variational information distillation, Bucket renormalization, Gauged variational inference, Gauged mini-bucket elimination, Odd-cycle matching BP, MCMC + belief propagation, Blossom belief propagation.
probabilistic objective
ID: probabilistic_objective.
Unsupervised combinatorial optimization.
progressive prediction
ID: progressive_prediction.
promoter DNA
ID: promoter_dna.
property prediction
ID: property_prediction.
Co-folding representations, HoliMol.
protein backbone
ID: protein_backbone.
protein engineering
ID: protein_engineering.
protein recognition
ID: protein_recognition.
Proteins & genomics
ID: proteins_genomics · Map category: d_bio.
Co-folding representations, TriProRep, VibeProteinBench, Riemannian MeanFlow, BioEmu-CV, DNAChunker, TPS-DPS, Adaptive Teachers, Antibody sequence–structure decoupling, Pessimistic backward policy, LED-GFN, Local Search GFlowNets, BootGen, RoMA.
protein–ligand co-folding
ID: proteinligand_co_folding.
proxy overestimation
ID: proxy_overestimation.
RoMA.
proxy-labeled data
ID: proxy_labeled_data.
pseudo-label
ID: pseudo_label.
Latent veracity inference, CORE-PO.
QH9
ID: qh9.
rank weighting
ID: rank_weighting.
ranking
ID: ranking.
Rare events & transition paths
ID: rare_events_transition_paths · Map category: c_rare_events.
RDKit
ID: rdkit.
reactant selection
ID: reactant_selection.
reaction template
ID: reaction_template.
reasoning chain
ID: reasoning_chain.
reasoning confidence
ID: reasoning_confidence.
Reinforcement learning
ID: reinforcement_learning · Map category: m_rl.
Search aliases: RL.
Co-folding representations, RL4CO, CORE-PO, Structurally diverse molecular LLMs, Symmetric replay training, Adaptive-grid exploration, Planner-guided imitation, Disentangled risk-sensitive MARL, Difficult, not too different, Learning what to defer.
relaxation
ID: relaxation.
Replay & off-policy reuse
ID: replay_off_policy_reuse · Map category: c_replay.
Search-guided diffusion samplers, TPS-DPS, Adaptive Teachers, Iterated energy-based flow matching, Symmetric replay training, Local Search GFlowNets, Genetic expert-guided learning.
Representation alignment
ID: representation_alignment · Map category: c_representation_alignment.
Co-folding representations, TriProRep.
Representation learning
ID: representation_learning · Map category: m_representation.
Co-folding representations, TriProRep, scTrilemma, BioEmu-CV, DNAChunker, CleanMol, Molecular Structural Reasoning, Antibody sequence–structure decoupling, HoliMol, BootGen, Concept intervention analysis, Planner-guided imitation, Learning with a biased committee, Difficult, not too different, Self-improved retrosynthesis, RetCL, Genetic expert-guided learning, Variational information distillation.
RepSP
ID: repsp.
residual edge
ID: residual_edge.
Spanning-tree molecular generation.
residue interaction
ID: residue_interaction.
retrosynthesis
ID: retrosynthesis.
reward-based optimization
ID: reward_based_optimization.
reward-proportional sampling
ID: reward_proportional_sampling.
rigid building block
ID: rigid_building_block.
risk sensitivity
ID: risk_sensitivity.
Disentangled risk-sensitive MARL.
RNA design
ID: rna_design.
Pessimistic backward policy, LED-GFN.
Robust learning
ID: robust_learning · Map category: m_robust.
scTrilemma, AdaPert, Multi-bias robust learning, Learning with a biased committee, Disentangled risk-sensitive MARL, RoMA, Learning from Failure.
rotation
ID: rotation.
rounding
ID: rounding.
Unsupervised combinatorial optimization.
routing
ID: routing.
Sample efficiency
ID: sample_efficiency · Map category: c_sample_efficiency.
Symmetric replay training, Planner-guided imitation, Learning what to defer.
sample reweighting
ID: sample_reweighting.
Scaling studies
ID: scaling_studies · Map category: c_scaling.
SCF initialization
ID: scf_initialization.
Search aliases: SCF; self-consistent field.
scientific debate
ID: scientific_debate.
score-based generation
ID: score_based_generation.
Search & optimization
ID: search_optimization · Map category: m_search.
MaskGXT / HACO, INDIBATOR, Latent veracity inference, Search-guided diffusion samplers, MT-Mol, RxnFlow, Symmetric replay training, Unsupervised combinatorial optimization, Adaptive-grid exploration, Local Search GFlowNets, BootGen, Planner-guided imitation, Spanning-tree molecular generation, RoMA, Self-improved retrosynthesis, RetCL, Genetic expert-guided learning, Learning what to defer, Odd-cycle matching BP, Blossom belief propagation.
Search–learning feedback
ID: searchlearning_feedback · Map category: c_search_learning.
Latent veracity inference, Search-guided diffusion samplers, Local Search GFlowNets, Planner-guided imitation, Self-improved retrosynthesis, Genetic expert-guided learning.
self-correction
ID: self_correction.
self-normalized importance sampling
ID: self_normalized_importance_sampling.
Energy-based generator matching.
self-supervised pretraining
ID: self_supervised_pretraining.
Learning with a biased committee.
Self-training
ID: self_training · Map category: c_self_training.
CORE-PO, BootGen, Self-improved retrosynthesis.
semantic similarity
ID: semantic_similarity.
separate flow times
ID: separate_flow_times.
set-autoregressive
ID: set_autoregressive.
Structurally diverse molecular LLMs.
side effects
ID: side_effects.
Signal–noise separation
ID: signalnoise_separation · Map category: c_signal_noise.
AdaPert, Multi-bias robust learning, Learning with a biased committee, Difficult, not too different, Learning from Failure.
Simulation / proxy-to-reality gap
ID: simulation_proxy_to_reality_gap · Map category: c_proxy_gap.
Search aliases: sim2real; simulation-to-real; proxy-to-reality.
Search alias for inspecting the validation gap; not a verified robotics sim-to-real claim.
Packora, MADField, VibeProteinBench, Structurally diverse molecular LLMs, BootGen, RoMA, Genetic expert-guided learning.
single-cell RNA
ID: single_cell_rna.
SkinCon
ID: skincon.
Concept intervention analysis.
slab–adsorbate
ID: slabadsorbate.
SMILES
ID: smiles.
SMILES parsing
ID: smiles_parsing.
SO(2)
ID: so_2.
solver initialization
ID: solver_initialization.
solver library
ID: solver_library.
spanning tree
ID: spanning_tree.
Spanning-tree molecular generation.
sparse adjacency
ID: sparse_adjacency.
spectral signal
ID: spectral_signal.
spherical feature grid
ID: spherical_feature_grid.
spurious correlation
ID: spurious_correlation.
step correctness
ID: step_correctness.
stochastic block model
ID: stochastic_block_model.
structural diversity
ID: structural_diversity.
Structurally diverse molecular LLMs.
structural probing
ID: structural_probing.
structural sketch
ID: structural_sketch.
Molecular Structural Reasoning.
structured prediction
ID: structured_prediction.
Structured tokenization
ID: structured_tokenization · Map category: c_tokenization.
TriProRep, MaskGXT / HACO, MELD, DNAChunker, CleanMol, Molecular Structural Reasoning, Gap-encoded edge lists, K²-tree graph generation, Spanning-tree molecular generation.
Student–Teacher
ID: studentteacher.
subgoal skipping
ID: subgoal_skipping.
subgraph parsing
ID: subgraph_parsing.
successful route imitation
ID: successful_route_imitation.
super-resolution
ID: super_resolution.
symmetric trajectories
ID: symmetric_trajectories.
Symmetry
ID: symmetry · Map category: c_symmetry.
MaskGXT / HACO, Multimodal Crystal Flow, QHFlow2, QHFlow, Symmetric replay training, Gauged variational inference, Gauged mini-bucket elimination.
symmetry token
ID: symmetry_token.
Synthesis & reactant availability
ID: synthesis_reactant_availability · Map category: c_synthesizability.
Search aliases: synthesizability.
RxnFlow, Self-improved retrosynthesis, RetCL.
teammate uncertainty
ID: teammate_uncertainty.
Disentangled risk-sensitive MARL.
tensor network
ID: tensor_network.
terminal energy decomposition
ID: terminal_energy_decomposition.
terminal look-ahead
ID: terminal_look_ahead.
terminal rewards
ID: terminal_rewards.
Time scales & dynamics
ID: time_scales_dynamics · Map category: c_time_scales.
time-lagged conditioning
ID: time_lagged_conditioning.
token repetition
ID: token_repetition.
Antibody sequence–structure decoupling.
Tool-grounded reasoning
ID: tool_grounded_reasoning · Map category: c_tool_grounding.
VibeProteinBench, INDIBATOR, PBio-Agent / LincsQA, MT-Mol.
torsion
ID: torsion.
trajectory collision
ID: trajectory_collision.
MELD.
trajectory refinement
ID: trajectory_refinement.
transition path sampling
ID: transition_path_sampling.
translation
ID: translation.
traveling salesman problem
ID: traveling_salesman_problem.
RL4CO, Symmetric replay training.
tree-aware Transformer
ID: tree_aware_transformer.
tree-relative position
ID: tree_relative_position.
Spanning-tree molecular generation.
two-stage pair update
ID: two_stage_pair_update.
unattained subgoal
ID: unattained_subgoal.
Uncertainty & risk
ID: uncertainty_risk · Map category: c_uncertainty.
Disentangled risk-sensitive MARL.
Unnormalized distributions
ID: unnormalized_distributions · Map category: c_unnormalized.
Energy-based generator matching, Iterated energy-based flow matching, Bucket renormalization, Gauged variational inference, Gauged mini-bucket elimination, MCMC + belief propagation.
unobserved backward flow
ID: unobserved_backward_flow.
unseen reaction template
ID: unseen_reaction_template.
upper bound
ID: upper_bound.
Gauged mini-bucket elimination.
UrbanCars
ID: urbancars.
USPTO
ID: uspto.
valence
ID: valence.
Spanning-tree molecular generation.
variable elimination
ID: variable_elimination.
variable-length tokens
ID: variable_length_tokens.
Variance reduction
ID: variance_reduction · Map category: c_variance_reduction.
Energy-based generator matching, TPS-DPS.
variational autoencoder
ID: variational_autoencoder.
Search aliases: VAE.
variational bound
ID: variational_bound.
Variational information distillation.
Veracity Search
ID: veracity_search.
Verification & diagnostics
ID: verification_diagnostics · Map category: c_verification.
Latent veracity inference, CORE-PO, CleanMol, Causal Influence Prompting, Concept intervention analysis.
voting
ID: voting.
weather
ID: weather.
Weather & cosmology
ID: weather_cosmology · Map category: d_geoscience.
weight rewinding
ID: weight_rewinding.
weighted mini-bucket
ID: weighted_mini_bucket.
Gauged mini-bucket elimination.
Keywords by paper
AtomMOF
AtomMOF: All-Atom Flow Matching for MOF-Adsorbate Structure Prediction · Work ID: kim2026atommof.
adsorbate (adsorbate); Adsorption & host–guest modeling (adsorption_hostguest_modeling); All-atom modeling (all_atom_modeling); Amortized inference / generation (amortized_inference_generation); carbon capture (carbon_capture); Diffusion Transformer (diffusion_transformer); Energy / reward guidance (energy_reward_guidance); Equivariant models (equivariant_models); Feynman–Kac steering (feynmankac_steering); Flow matching (flow_matching); interatomic potential (interatomic_potential); Materials & electronic structure (materials_electronic_structure); MOF (mof); Molecules & drug discovery (molecules_drug_discovery); Scaling studies (scaling_studies).
Co-folding representations
A Systematic Evaluation of Co-folding Model Representations for Small-Molecule Learning · Work ID: jang2026boltz.
ADMET (admet); Benchmarking (benchmarking); Benchmarks & evaluation (benchmarks_evaluation); Boltz2 (boltz2); Credit assignment (credit_assignment); Cross-modal learning (cross_modal_learning); Knowledge distillation (knowledge_distillation); Molecules & drug discovery (molecules_drug_discovery); property prediction (property_prediction); protein–ligand co-folding (proteinligand_co_folding); Proteins & genomics (proteins_genomics); Reinforcement learning (reinforcement_learning); Representation alignment (representation_alignment); Representation learning (representation_learning); reward-based optimization (reward_based_optimization).
TriProRep
Atom-level Protein Representation Learning Improves Protein Structure Prediction · Work ID: kim2026atomlevel.
All-atom modeling (all_atom_modeling); Benchmarking (benchmarking); Benchmarks & evaluation (benchmarks_evaluation); Cross-modal learning (cross_modal_learning); full-atom token (full_atom_token); homodimer (homodimer); Language / sequence models (language_sequence_models); Proteins & genomics (proteins_genomics); Representation alignment (representation_alignment); Representation learning (representation_learning); RepSP (repsp); residue interaction (residue_interaction); structural probing (structural_probing); Structured tokenization (structured_tokenization).
scTrilemma
scTrilemma: Balancing Identity, Invariance, and Reconstruction in Single-Cell Representation Learning · Work ID: oh2026sctrilemma.
batch effects (batch_effects); Benchmarking (benchmarking); Cells & perturbations (cells_perturbations); Competing objectives (competing_objectives); context-conditioned prior (context_conditioned_prior); Information routing (information_routing); Invariance & spurious factors (invariance_spurious_factors); latent bottleneck (latent_bottleneck); Probabilistic inference (probabilistic_inference); Representation learning (representation_learning); Robust learning (robust_learning); single-cell RNA (single_cell_rna); variational autoencoder (variational_autoencoder).
Packora
Packora: Systematic Design for Generative Molecular Crystal Structure Prediction · Work ID: kim2026packora.
All-atom modeling (all_atom_modeling); Benchmarking (benchmarking); Benchmarks & evaluation (benchmarks_evaluation); blind test (blind_test); Conditional generation (conditional_generation); Equivariant models (equivariant_models); Flow matching (flow_matching); Materials & electronic structure (materials_electronic_structure); molecular crystal (molecular_crystal); Molecules & drug discovery (molecules_drug_discovery); polymorph (polymorph); ranking (ranking); relaxation (relaxation); Scaling studies (scaling_studies); Simulation / proxy-to-reality gap (simulation_proxy_to_reality_gap).
MaskGXT / HACO
Discovering Crystal Structure Prediction Algorithms with an AI Co-Scientist · Work ID: seong2026discovering.
Algorithm discovery (algorithm_discovery); Benchmarking (benchmarking); Diversity & mode coverage (diversity_mode_coverage); HACO (haco); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); masked generative modeling (masked_generative_modeling); MaskGXT (maskgxt); Materials & electronic structure (materials_electronic_structure); Method transfer (method_transfer); polymorph coverage (polymorph_coverage); Search & optimization (search_optimization); Structured tokenization (structured_tokenization); Symmetry (symmetry); symmetry token (symmetry_token).
MADField
MADField: Multi-fidelity Amortized Density Field for Adsorption in Nanoporous Materials · Work ID: kim2026madfield.
Adsorption & host–guest modeling (adsorption_hostguest_modeling); Amortized inference / generation (amortized_inference_generation); ARC-MOF (arc_mof); cDFT (cdft); Density / field representations (density_field_representations); gas uptake (gas_uptake); GCMC (gcmc); Learned / classical hybrids (learned_classical_hybrids); Materials & electronic structure (materials_electronic_structure); Multi-fidelity learning (multi_fidelity_learning); Neural fields & operators (neural_fields_operators); Simulation / proxy-to-reality gap (simulation_proxy_to_reality_gap); solver initialization (solver_initialization).
VibeProteinBench
VibeProteinBench: An Evaluation Benchmark for Language-interfaced Vibe Protein Design · Work ID: seo2026vibeproteinbench.
Benchmarking (benchmarking); Benchmarks & evaluation (benchmarks_evaluation); Cross-modal learning (cross_modal_learning); expert rationale (expert_rationale); in silico validation (in_silico_validation); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); natural-language design (natural_language_design); protein engineering (protein_engineering); protein recognition (protein_recognition); Proteins & genomics (proteins_genomics); Simulation / proxy-to-reality gap (simulation_proxy_to_reality_gap); Tool-grounded reasoning (tool_grounded_reasoning).
AdaPert
Learning Adaptive Perturbation-Conditioned Contexts for Robust Transcriptional Response Prediction · Work ID: piao2026learning.
Benchmarking (benchmarking); cell-line transfer (cell_line_transfer); Cells & perturbations (cells_perturbations); Context selection (context_selection); CRISPR (crispr); differential expression (differential_expression); Graph neural networks (graph_neural_networks); Graph structure & transformations (graph_structure_transformations); knowledge graph (knowledge_graph); perturbation response (perturbation_response); Robust learning (robust_learning); Signal–noise separation (signalnoise_separation).
INDIBATOR
INDIBATOR: Diverse and Fact-Grounded Individuality for Multi-Agent Debate in Molecular Discovery · Work ID: jang2026indibator.
agent profiles (agent_profiles); bioactivity (bioactivity); Diversity & mode coverage (diversity_mode_coverage); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); lead optimization (lead_optimization); Molecules & drug discovery (molecules_drug_discovery); Multi-agent deliberation (multi_agent_deliberation); scientific debate (scientific_debate); Search & optimization (search_optimization); Tool-grounded reasoning (tool_grounded_reasoning); voting (voting).
Multimodal Crystal Flow
Multimodal Crystal Flow: Any-to-Any Modality Generation for Unified Crystal Modeling · Work ID: seong2026multimodal.
any-to-any generation (any_to_any_generation); composition ordering (composition_ordering); Conditional generation (conditional_generation); Cross-modal learning (cross_modal_learning); crystal structure prediction (crystal_structure_prediction); Flow matching (flow_matching); Materials & electronic structure (materials_electronic_structure); separate flow times (separate_flow_times); Symmetry (symmetry).
PBio-Agent / LincsQA
Progressive Multi-Agent Reasoning for Biological Perturbation Prediction · Work ID: kim2026progressive.
Benchmarking (benchmarking); Benchmarks & evaluation (benchmarks_evaluation); Cells & perturbations (cells_perturbations); Context selection (context_selection); Curriculum & training difficulty (curriculum_training_difficulty); gene regulation (gene_regulation); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); LincsQA (lincsqa); mechanism of action (mechanism_of_action); Multi-agent deliberation (multi_agent_deliberation); PerturbQA (perturbqa); progressive prediction (progressive_prediction); Tool-grounded reasoning (tool_grounded_reasoning).
Riemannian MeanFlow
Riemannian MeanFlow · Work ID: woo2026riemannian.
average velocity (average_velocity); Energy / reward guidance (energy_reward_guidance); Equivariant models (equivariant_models); Few-step generation (few_step_generation); flow map (flow_map); Flow matching (flow_matching); General learning & inference (general_learning_inference); Manifold geometry (manifold_geometry); promoter DNA (promoter_dna); protein backbone (protein_backbone); Proteins & genomics (proteins_genomics); terminal look-ahead (terminal_look_ahead).
QHFlow2
Machine Learning Hamiltonians are Accurate Energy-Force Predictors · Work ID: kim2026machine.
Benchmarking (benchmarking); Benchmarks & evaluation (benchmarks_evaluation); electronic Hamiltonian (electronic_hamiltonian); energy and forces (energy_and_forces); Equivariant models (equivariant_models); Graph neural networks (graph_neural_networks); Materials & electronic structure (materials_electronic_structure); Molecules & drug discovery (molecules_drug_discovery); Physical observables (physical_observables); QH9 (qh9); Scaling studies (scaling_studies); SO(2) (so_2); Symmetry (symmetry); two-stage pair update (two_stage_pair_update).
CatFlow
CatFlow: Co-generation of Slab-Adsorbate Systems via Flow Matching · Work ID: kim2026catflow.
Adsorption & host–guest modeling (adsorption_hostguest_modeling); catalyst (catalyst); Conditional generation (conditional_generation); Equivariant models (equivariant_models); Factorization & decomposition (factorization_decomposition); Flow matching (flow_matching); Materials & electronic structure (materials_electronic_structure); Molecules & drug discovery (molecules_drug_discovery); OC20 (oc20); orientation (orientation); Physics-informed priors (physics_informed_priors); primitive cell (primitive_cell); slab–adsorbate (slabadsorbate).
Latent veracity inference
Latent Veracity Inference for Identifying Errors in Stepwise Reasoning · Work ID: kim2026latent.
Amortized inference / generation (amortized_inference_generation); General learning & inference (general_learning_inference); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); Local feedback (local_feedback); Probabilistic inference (probabilistic_inference); pseudo-label (pseudo_label); reasoning chain (reasoning_chain); Search & optimization (search_optimization); Search–learning feedback (searchlearning_feedback); self-correction (self_correction); step correctness (step_correctness); Veracity Search (veracity_search); Verification & diagnostics (verification_diagnostics).
MELD
Learning Flexible Forward Trajectories for Masked Molecular Diffusion · Work ID: seo2026learning.
Adaptive trajectories (adaptive_trajectories); Constraint-aware design (constraint_aware_design); corruption schedule (corruption_schedule); Diffusion models (diffusion_models); Graph neural networks (graph_neural_networks); masked diffusion (masked_diffusion); molecular validity (molecular_validity); Molecules & drug discovery (molecules_drug_discovery); polymer (polymer); Structured tokenization (structured_tokenization); trajectory collision (trajectory_collision).
BioEmu-CV
Learning Collective Variables from BioEmu with Time-Lagged Generation · Work ID: park2026learning.
Benchmarking (benchmarking); BioEmu (bioemu); collective variables (collective_variables); Conditional generation (conditional_generation); Diffusion models (diffusion_models); enhanced sampling (enhanced_sampling); free energy (free_energy); Learned / classical hybrids (learned_classical_hybrids); Proteins & genomics (proteins_genomics); Rare events & transition paths (rare_events_transition_paths); Representation learning (representation_learning); time-lagged conditioning (time_lagged_conditioning); Time scales & dynamics (time_scales_dynamics).
DNAChunker
DNACHUNKER: Learnable Tokenization for DNA Language Models · Work ID: kim2026dnachunker.
Adaptive resolution (adaptive_resolution); Compression (compression); genomic segmentation (genomic_segmentation); Language / sequence models (language_sequence_models); masked language model (masked_language_model); mutation resilience (mutation_resilience); Proteins & genomics (proteins_genomics); Representation learning (representation_learning); Structured tokenization (structured_tokenization); variable-length tokens (variable_length_tokens).
RL4CO
RL4CO: an Extensive Reinforcement Learning for Combinatorial Optimization Benchmark · Work ID: berto2025rlco.
Benchmarking (benchmarking); Benchmarks & evaluation (benchmarks_evaluation); Combinatorial optimization (combinatorial_optimization); Graphs & discrete problems (graphs_discrete_problems); matched sample budgets (matched_sample_budgets); Modular frameworks (modular_frameworks); Reinforcement learning (reinforcement_learning); routing (routing); solver library (solver_library); traveling salesman problem (traveling_salesman_problem).
Energy-based generator matching
Energy-based Generator Matching: A Neural Sampler for General State Space · Work ID: woo2025energybased.
Amortized inference / generation (amortized_inference_generation); bootstrapping (bootstrapping); Diffusion models (diffusion_models); Discrete–continuous states (discretecontinuous_states); Energy-based sampling (energy_based_sampling); Flow matching (flow_matching); General learning & inference (general_learning_inference); generator matching (generator_matching); Importance sampling (importance_sampling); jump process (jump_process); self-normalized importance sampling (self_normalized_importance_sampling); Unnormalized distributions (unnormalized_distributions); Variance reduction (variance_reduction).
Search-guided diffusion samplers
On Scalable and Efficient Training of Diffusion Samplers · Work ID: kim2025scalable.
Amortized inference / generation (amortized_inference_generation); conformer (conformer); diffusion Learner (diffusion_learner); Diffusion models (diffusion_models); Diversity & mode coverage (diversity_mode_coverage); Energy-based sampling (energy_based_sampling); Exploration–exploitation (explorationexploitation); General learning & inference (general_learning_inference); MCMC (mcmc); MCMC Searcher (mcmc_searcher); Molecules & drug discovery (molecules_drug_discovery); novelty reward (novelty_reward); primacy bias (primacy_bias); Replay & off-policy reuse (replay_off_policy_reuse); Search & optimization (search_optimization); Search–learning feedback (searchlearning_feedback).
MOFFlow-2
Flexible MOF Generation with Torsion-Aware Flow Matching · Work ID: kim2025flexible.
Conditional generation (conditional_generation); Constraint-aware design (constraint_aware_design); Equivariant models (equivariant_models); Factorization & decomposition (factorization_decomposition); flexible building block (flexible_building_block); Flow matching (flow_matching); Language / sequence models (language_sequence_models); lattice (lattice); Manifold geometry (manifold_geometry); Materials & electronic structure (materials_electronic_structure); metal library (metal_library); Molecules & drug discovery (molecules_drug_discovery); SMILES (smiles); torsion (torsion).
QHFlow
High-order Equivariant Flow Matching for Density Functional Theory Hamiltonian Prediction · Work ID: kim2025highorder.
Conditional generation (conditional_generation); electronic Hamiltonian (electronic_hamiltonian); Equivariant models (equivariant_models); Flow matching (flow_matching); Learned / classical hybrids (learned_classical_hybrids); Materials & electronic structure (materials_electronic_structure); Molecules & drug discovery (molecules_drug_discovery); orbital alignment (orbital_alignment); Physical observables (physical_observables); QH9 (qh9); SCF initialization (scf_initialization); Symmetry (symmetry).
CORE-PO
Self-Training Large Language Models with Confident Reasoning · Work ID: jang2025selftraining.
accidental correct answers (accidental_correct_answers); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); Local feedback (local_feedback); policy optimization (policy_optimization); pseudo-label (pseudo_label); reasoning confidence (reasoning_confidence); Reinforcement learning (reinforcement_learning); Self-training (self_training); Verification & diagnostics (verification_diagnostics).
MT-Mol
MT-Mol: Multi Agent System with Tool-based Reasoning for Molecular Optimization · Work ID: kim2025mtmol.
chemist agents (chemist_agents); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); Local feedback (local_feedback); low-budget optimization (low_budget_optimization); Molecules & drug discovery (molecules_drug_discovery); Multi-agent deliberation (multi_agent_deliberation); PMO-1K (pmo_1k); RDKit (rdkit); Search & optimization (search_optimization); Tool-grounded reasoning (tool_grounded_reasoning).
CleanMol
Improving Chemical Understanding of LLMs via SMILES Parsing · Work ID: jang2025improving.
Curriculum & training difficulty (curriculum_training_difficulty); global graph (global_graph); Graph structure & transformations (graph_structure_transformations); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); Mol-Instructions (mol_instructions); Molecules & drug discovery (molecules_drug_discovery); Representation learning (representation_learning); SMILES parsing (smiles_parsing); Structured tokenization (structured_tokenization); subgraph parsing (subgraph_parsing); Verification & diagnostics (verification_diagnostics).
Causal Influence Prompting
Enhancing LLM Agent Safety via Causal Influence Prompting · Work ID: hahm2025enhancing.
agent safety (agent_safety); causal influence diagram (causal_influence_diagram); Causal reasoning & interventions (causal_reasoning_interventions); Context selection (context_selection); decision utilities (decision_utilities); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); side effects (side_effects); Verification & diagnostics (verification_diagnostics).
Molecular Structural Reasoning
Structural Reasoning Improves Molecular Understanding of LLM · Work ID: jang2025structural.
Graph structure & transformations (graph_structure_transformations); intermediate reasoning (intermediate_reasoning); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); Local feedback (local_feedback); molecular graph reconstruction (molecular_graph_reconstruction); Molecules & drug discovery (molecules_drug_discovery); Representation learning (representation_learning); structural sketch (structural_sketch); Structured tokenization (structured_tokenization).
RxnFlow
Generative Flows on Synthetic Pathway for Drug Design · Work ID: seo2025generative.
action subsampling (action_subsampling); building-block library (building_block_library); Combinatorial optimization (combinatorial_optimization); Conditional generation (conditional_generation); Constraint-aware design (constraint_aware_design); Diversity & mode coverage (diversity_mode_coverage); docking (docking); GFlowNets (gflownets); Molecules & drug discovery (molecules_drug_discovery); pocket conditioning (pocket_conditioning); reaction template (reaction_template); Search & optimization (search_optimization); Synthesis & reactant availability (synthesis_reactant_availability).
MOFFlow
MOFFlow: Flow Matching for Structure Prediction of Metal-Organic Frameworks · Work ID: kim2025mofflow.
Amortized inference / generation (amortized_inference_generation); Conditional generation (conditional_generation); Equivariant models (equivariant_models); Factorization & decomposition (factorization_decomposition); Flow matching (flow_matching); lattice (lattice); Manifold geometry (manifold_geometry); Materials & electronic structure (materials_electronic_structure); MOF (mof); rigid building block (rigid_building_block); rotation (rotation); translation (translation).
TPS-DPS
Transition Path Sampling with Improved Off-Policy Training of Diffusion Path Samplers · Work ID: seong2025transition.
alanine dipeptide (alanine_dipeptide); Amortized inference / generation (amortized_inference_generation); control variates (control_variates); Diffusion models (diffusion_models); Energy-based sampling (energy_based_sampling); equivariant bias (equivariant_bias); Equivariant models (equivariant_models); log-variance loss (log_variance_loss); Molecules & drug discovery (molecules_drug_discovery); Proteins & genomics (proteins_genomics); Rare events & transition paths (rare_events_transition_paths); Replay & off-policy reuse (replay_off_policy_reuse); Time scales & dynamics (time_scales_dynamics); transition path sampling (transition_path_sampling); Variance reduction (variance_reduction).
ReBind
ReBind: Enhancing Ground-state Molecular Conformation Prediction via Force-Based Graph Rewiring · Work ID: kim2025rebind.
All-atom modeling (all_atom_modeling); conformation (conformation); Equivariant models (equivariant_models); Graph neural networks (graph_neural_networks); graph rewiring (graph_rewiring); Graph structure & transformations (graph_structure_transformations); Lennard-Jones (lennard_jones); low-degree atoms (low_degree_atoms); Molecules & drug discovery (molecules_drug_discovery); non-bonded interactions (non_bonded_interactions); Physics-informed priors (physics_informed_priors).
Adaptive Teachers
Adaptive Teachers for Amortized Samplers · Work ID: kim2025adaptive.
Amortized inference / generation (amortized_inference_generation); Auxiliary teachers (auxiliary_teachers); continuous sampling (continuous_sampling); Curriculum & training difficulty (curriculum_training_difficulty); Diversity & mode coverage (diversity_mode_coverage); Energy-based sampling (energy_based_sampling); Exploration–exploitation (explorationexploitation); General learning & inference (general_learning_inference); GFlowNets (gflownets); high-loss regions (high_loss_regions); mode coverage (mode_coverage); Molecules & drug discovery (molecules_drug_discovery); Proteins & genomics (proteins_genomics); Replay & off-policy reuse (replay_off_policy_reuse); Student–Teacher (studentteacher).
Antibody sequence–structure decoupling
Decoupled Sequence and Structure Generation for Realistic Antibody Design · Work ID: kim2024decoupled.
affinity proxy (affinity_proxy); antibody (antibody); CDR-H3 (cdr_h3); Competing objectives (competing_objectives); composition objective (composition_objective); Constraint-aware design (constraint_aware_design); Diversity & mode coverage (diversity_mode_coverage); Equivariant models (equivariant_models); Factorization & decomposition (factorization_decomposition); Proteins & genomics (proteins_genomics); Representation learning (representation_learning); token repetition (token_repetition).
Structurally diverse molecular LLMs
Can LLMs Generate Diverse Molecules? Towards Alignment with Structural Diversity · Work ID: jang2024can.
Conditional generation (conditional_generation); description-guided generation (description_guided_generation); Diversity & mode coverage (diversity_mode_coverage); Language & scientific reasoning (language_scientific_reasoning); Language / sequence models (language_sequence_models); Molecules & drug discovery (molecules_drug_discovery); Reinforcement learning (reinforcement_learning); set-autoregressive (set_autoregressive); Simulation / proxy-to-reality gap (simulation_proxy_to_reality_gap); structural diversity (structural_diversity).
Iterated energy-based flow matching
Iterated Energy-based Flow Matching for Sampling from Boltzmann Densities · Work ID: woo2024iterated.
Amortized inference / generation (amortized_inference_generation); double well (double_well); Energy-based sampling (energy_based_sampling); Flow matching (flow_matching); Gaussian mixture (gaussian_mixture); General learning & inference (general_learning_inference); Importance sampling (importance_sampling); off-policy flow matching (off_policy_flow_matching); partition function (partition_function); Replay & off-policy reuse (replay_off_policy_reuse); Unnormalized distributions (unnormalized_distributions).
Non-backtracking GNNs
Non-backtracking Graph Neural Networks · Work ID: park2024nonbacktracking.
expressivity (expressivity); Graph neural networks (graph_neural_networks); Graph structure & transformations (graph_structure_transformations); Graphs & discrete problems (graphs_discrete_problems); Long-range propagation (long_range_propagation); Message passing (message_passing); node sensitivity (node_sensitivity); non-backtracking (non_backtracking); oversquashing (oversquashing); stochastic block model (stochastic_block_model).
Pessimistic backward policy
Pessimistic Backward Policy for GFlowNets · Work ID: jang2024pessimistic.
Adaptive trajectories (adaptive_trajectories); Amortized inference / generation (amortized_inference_generation); backward policy (backward_policy); Diversity & mode coverage (diversity_mode_coverage); Exploration–exploitation (explorationexploitation); General learning & inference (general_learning_inference); GFlowNets (gflownets); Molecules & drug discovery (molecules_drug_discovery); Proteins & genomics (proteins_genomics); RNA design (rna_design); terminal rewards (terminal_rewards); unobserved backward flow (unobserved_backward_flow).
Spherical neural fields
Hybrid Neural Representation for Spherical Data · Work ID: kim2024hybrid.
Adaptive resolution (adaptive_resolution); Compression (compression); cosmic microwave background (cosmic_microwave_background); Density / field representations (density_field_representations); HNeR (hner); Manifold geometry (manifold_geometry); Neural fields & operators (neural_fields_operators); spherical feature grid (spherical_feature_grid); super-resolution (super_resolution); weather (weather); Weather & cosmology (weather_cosmology).
Gaussian plane-wave neural operator
Gaussian Plane-Wave Neural Operator for Electron Density Estimation · Work ID: kim2024gaussian.
Density / field representations (density_field_representations); electron density (electron_density); Equivariant models (equivariant_models); exchange-correlation (exchange_correlation); Factorization & decomposition (factorization_decomposition); frequency decomposition (frequency_decomposition); Gaussian orbital (gaussian_orbital); Materials & electronic structure (materials_electronic_structure); Molecules & drug discovery (molecules_drug_discovery); Neural fields & operators (neural_fields_operators); Physical observables (physical_observables); Physics-informed priors (physics_informed_priors); plane wave (plane_wave).
Multi-bias robust learning
Improving Robustness to Multiple Spurious Correlations by Multi-Objective Optimization · Work ID: kim2024improving.
Benchmarking (benchmarking); bias attributes (bias_attributes); Competing objectives (competing_objectives); General learning & inference (general_learning_inference); group loss (group_loss); Invariance & spurious factors (invariance_spurious_factors); minimax Pareto (minimax_pareto); MultiCelebA (multiceleba); Robust learning (robust_learning); Signal–noise separation (signalnoise_separation); UrbanCars (urbancars).
Symmetric replay training
Enhancing Sample Efficiency in Black-box Combinatorial Optimization via Symmetric Replay Training · Work ID: kim2024enhancing.
Combinatorial optimization (combinatorial_optimization); Exploration–exploitation (explorationexploitation); Graphs & discrete problems (graphs_discrete_problems); hardware design (hardware_design); Molecules & drug discovery (molecules_drug_discovery); oracle budget (oracle_budget); Reinforcement learning (reinforcement_learning); Replay & off-policy reuse (replay_off_policy_reuse); Sample efficiency (sample_efficiency); Search & optimization (search_optimization); symmetric trajectories (symmetric_trajectories); Symmetry (symmetry); traveling salesman problem (traveling_salesman_problem).
Unsupervised combinatorial optimization
Tackling Complex Conditions in Unsupervised Combinatorial Optimization · Work ID: bu2024tackling.
cardinality (cardinality); Combinatorial optimization (combinatorial_optimization); Constraint-aware design (constraint_aware_design); Derandomization (derandomization); facility location (facility_location); Graphs & discrete problems (graphs_discrete_problems); maximum coverage (maximum_coverage); Probabilistic inference (probabilistic_inference); probabilistic objective (probabilistic_objective); rounding (rounding); Search & optimization (search_optimization).
Adaptive-grid exploration
Breadth-First Exploration in Adaptive Grid-based Reinforcement Learning · Work ID: yoon2024breadthfirst.
adaptive grid (adaptive_grid); Adaptive resolution (adaptive_resolution); breadth-first exploration (breadth_first_exploration); Constraint-aware design (constraint_aware_design); Control & multi-agent learning (control_multi_agent_learning); Exploration–exploitation (explorationexploitation); goal-conditioned RL (goal_conditioned_rl); Graph structure & transformations (graph_structure_transformations); Reinforcement learning (reinforcement_learning); Search & optimization (search_optimization); unattained subgoal (unattained_subgoal).
HoliMol
Holistic Molecular Representation Learning via Multi-view Fragmentation · Work ID: kim2024holistic.
2D–3D views (views_2d3d); Cross-modal learning (cross_modal_learning); Factorization & decomposition (factorization_decomposition); fragment contrast (fragment_contrast); Graph neural networks (graph_neural_networks); molecular fragmentation (molecular_fragmentation); Molecules & drug discovery (molecules_drug_discovery); property prediction (property_prediction); Representation learning (representation_learning).
EPIC
EPIC: Graph Augmentation with Edit Path Interpolation via Learnable Cost · Work ID: heo2024epic.
Adaptive trajectories (adaptive_trajectories); Constraint-aware design (constraint_aware_design); context-sensitive cost (context_sensitive_cost); Data augmentation (data_augmentation); graph edit path (graph_edit_path); Graph neural networks (graph_neural_networks); Graph structure & transformations (graph_structure_transformations); Graphs & discrete problems (graphs_discrete_problems); interpolation (interpolation); label corruption (label_corruption).
LED-GFN
Learning Energy Decompositions for Partial Inference in GFlowNets · Work ID: jang2024learning.
Amortized inference / generation (amortized_inference_generation); Credit assignment (credit_assignment); Factorization & decomposition (factorization_decomposition); General learning & inference (general_learning_inference); GFlowNets (gflownets); Graphs & discrete problems (graphs_discrete_problems); Local feedback (local_feedback); maximum independent set (maximum_independent_set); Molecules & drug discovery (molecules_drug_discovery); potential function (potential_function); Probabilistic inference (probabilistic_inference); Proteins & genomics (proteins_genomics); RNA design (rna_design); terminal energy decomposition (terminal_energy_decomposition).
Gap-encoded edge lists
A Simple and Scalable Representation for Graph Generation · Work ID: jang2024simple.
attributed grammar (attributed_grammar); bandwidth (bandwidth); Compression (compression); Compression & compact coding (compression_compact_coding); edge list (edge_list); gap encoding (gap_encoding); GEEL (geel); Graph structure & transformations (graph_structure_transformations); Graphs & discrete problems (graphs_discrete_problems); Language / sequence models (language_sequence_models); Molecules & drug discovery (molecules_drug_discovery); Structured tokenization (structured_tokenization).
K²-tree graph generation
Graph Generation with K^2 Trees · Work ID: jang2024graph.
Adaptive resolution (adaptive_resolution); Compression (compression); Compression & compact coding (compression_compact_coding); Graph structure & transformations (graph_structure_transformations); Graphs & discrete problems (graphs_discrete_problems); hierarchical tokens (hierarchical_tokens); K²-tree (k2_tree); Language / sequence models (language_sequence_models); Molecules & drug discovery (molecules_drug_discovery); sparse adjacency (sparse_adjacency); Structured tokenization (structured_tokenization); tree-aware Transformer (tree_aware_transformer).
Local Search GFlowNets
Local Search GFlowNets · Work ID: kim2024local.
backtracking (backtracking); Diversity & mode coverage (diversity_mode_coverage); Exploration–exploitation (explorationexploitation); forward reconstruction (forward_reconstruction); General learning & inference (general_learning_inference); GFlowNets (gflownets); Local search (local_search); Molecules & drug discovery (molecules_drug_discovery); Proteins & genomics (proteins_genomics); Replay & off-policy reuse (replay_off_policy_reuse); reward-proportional sampling (reward_proportional_sampling); Search & optimization (search_optimization); Search–learning feedback (searchlearning_feedback); trajectory refinement (trajectory_refinement).
Wavelet graph diffusion
Multi-resolution Spectral Coherence for Graph Generation with Score-based Diffusion · Work ID: cho2023multiresolution.
Adaptive resolution (adaptive_resolution); Density / field representations (density_field_representations); Diffusion models (diffusion_models); Graph neural networks (graph_neural_networks); Graph structure & transformations (graph_structure_transformations); graph wavelet (graph_wavelet); Graphs & discrete problems (graphs_discrete_problems); node–edge coupling (nodeedge_coupling); score-based generation (score_based_generation); spectral signal (spectral_signal).
Structured node diffusion
Diffusion Probabilistic Models for Structured Node Classification · Work ID: jang2023diffusion.
AGG-WL (agg_wl); Conditional generation (conditional_generation); constrained sampling (constrained_sampling); Constraint-aware design (constraint_aware_design); Diffusion models (diffusion_models); Graph neural networks (graph_neural_networks); Graphs & discrete problems (graphs_discrete_problems); Message passing (message_passing); partially observed labels (partially_observed_labels); Probabilistic inference (probabilistic_inference); structured prediction (structured_prediction).
BootGen
Bootstrapped Training of Score-Conditioned Generator for Offline Design of Biological Sequences · Work ID: kim2023bootstrapped.
Amortized inference / generation (amortized_inference_generation); Diversity & mode coverage (diversity_mode_coverage); Knowledge distillation (knowledge_distillation); model aggregation (model_aggregation); offline design (offline_design); Proteins & genomics (proteins_genomics); proxy-labeled data (proxy_labeled_data); rank weighting (rank_weighting); Representation learning (representation_learning); Search & optimization (search_optimization); Self-training (self_training); Simulation / proxy-to-reality gap (simulation_proxy_to_reality_gap).
Concept intervention analysis
A Closer Look at the Intervention Procedure of Concept Bottleneck Models · Work ID: shin2023closer.
Benchmarking (benchmarking); Benchmarks & evaluation (benchmarks_evaluation); Causal reasoning & interventions (causal_reasoning_interventions); concept bottleneck (concept_bottleneck); CUB (cub); fairness (fairness); General learning & inference (general_learning_inference); intervention budget (intervention_budget); Local feedback (local_feedback); Representation learning (representation_learning); SkinCon (skincon); Verification & diagnostics (verification_diagnostics).
Planner-guided imitation
Imitating Graph-Based Planning with Goal-Conditioned Policies · Work ID: kim2023imitating.
Control & multi-agent learning (control_multi_agent_learning); goal-conditioned policy (goal_conditioned_policy); graph planner (graph_planner); Knowledge distillation (knowledge_distillation); planner-free execution (planner_free_execution); Reinforcement learning (reinforcement_learning); Representation learning (representation_learning); Sample efficiency (sample_efficiency); Search & optimization (search_optimization); Search–learning feedback (searchlearning_feedback); subgoal skipping (subgoal_skipping).
Learning with a biased committee
Learning Debiased Classifier with Biased Committee · Work ID: kim2022learning.
Auxiliary teachers (auxiliary_teachers); bias-conflicting samples (bias_conflicting_samples); biased committee (biased_committee); Curriculum & training difficulty (curriculum_training_difficulty); General learning & inference (general_learning_inference); Invariance & spurious factors (invariance_spurious_factors); knowledge exchange (knowledge_exchange); Representation learning (representation_learning); Robust learning (robust_learning); self-supervised pretraining (self_supervised_pretraining); Signal–noise separation (signalnoise_separation).
Disentangled risk-sensitive MARL
Disentangling Sources of Risk for Distributional Multi-Agent Reinforcement Learning · Work ID: son2022disentangling.
Control & multi-agent learning (control_multi_agent_learning); distributional value (distributional_value); environment uncertainty (environment_uncertainty); Exploration–exploitation (explorationexploitation); Factorization & decomposition (factorization_decomposition); Reinforcement learning (reinforcement_learning); risk sensitivity (risk_sensitivity); Robust learning (robust_learning); teammate uncertainty (teammate_uncertainty); Uncertainty & risk (uncertainty_risk).
Difficult, not too different
What Makes Better Augmentation Strategies? Augment Difficult but Not Too Different · Work ID: kim2022what.
Constraint-aware design (constraint_aware_design); Curriculum & training difficulty (curriculum_training_difficulty); Data augmentation (data_augmentation); GLUE (glue); Language & scientific reasoning (language_scientific_reasoning); learned augmentation (learned_augmentation); low confidence (low_confidence); Reinforcement learning (reinforcement_learning); Representation learning (representation_learning); sample reweighting (sample_reweighting); semantic similarity (semantic_similarity); Signal–noise separation (signalnoise_separation).
Spanning-tree molecular generation
Spanning Tree-based Graph Generation for Molecules · Work ID: ahn2022spanning.
compact tree operation (compact_tree_operation); Constraint-aware design (constraint_aware_design); Graph structure & transformations (graph_structure_transformations); Graphs & discrete problems (graphs_discrete_problems); Language / sequence models (language_sequence_models); Molecules & drug discovery (molecules_drug_discovery); residual edge (residual_edge); Search & optimization (search_optimization); spanning tree (spanning_tree); Structured tokenization (structured_tokenization); tree-relative position (tree_relative_position); valence (valence).
RoMA
RoMA: Robust Model Adaptation for Offline Model-based Optimization · Work ID: yu2021roma.
Constraint-aware design (constraint_aware_design); Control & multi-agent learning (control_multi_agent_learning); Design-bench (design_bench); General learning & inference (general_learning_inference); Local search (local_search); local smoothness (local_smoothness); offline model-based optimization (offline_model_based_optimization); Proteins & genomics (proteins_genomics); proxy overestimation (proxy_overestimation); Robust learning (robust_learning); Search & optimization (search_optimization); Simulation / proxy-to-reality gap (simulation_proxy_to_reality_gap).
Self-improved retrosynthesis
Self-Improved Retrosynthetic Planning · Work ID: kim2021selfimproved.
available reactants (available_reactants); Constraint-aware design (constraint_aware_design); Data augmentation (data_augmentation); forward reaction augmentation (forward_reaction_augmentation); Molecules & drug discovery (molecules_drug_discovery); Representation learning (representation_learning); retrosynthesis (retrosynthesis); Search & optimization (search_optimization); Search–learning feedback (searchlearning_feedback); Self-training (self_training); successful route imitation (successful_route_imitation); Synthesis & reactant availability (synthesis_reactant_availability).
RetCL
RetCL: A Selection-based Approach for Retrosynthesis via Contrastive Learning · Work ID: lee2021retcl.
candidate inventory (candidate_inventory); Constraint-aware design (constraint_aware_design); Contrastive learning (contrastive_learning); Graph neural networks (graph_neural_networks); hard negatives (hard_negatives); Molecules & drug discovery (molecules_drug_discovery); reactant selection (reactant_selection); Representation learning (representation_learning); Search & optimization (search_optimization); Synthesis & reactant availability (synthesis_reactant_availability); unseen reaction template (unseen_reaction_template); USPTO (uspto).
Layer-adaptive pruning
Layer-adaptive sparsity for the Magnitude-based Pruning · Work ID: lee2021layeradaptive.
Adaptive resolution (adaptive_resolution); Compression (compression); Compression & compact coding (compression_compact_coding); General learning & inference (general_learning_inference); layerwise sparsity (layerwise_sparsity); magnitude pruning (magnitude_pruning); model distortion (model_distortion); weight rewinding (weight_rewinding).
Learning from Failure
Learning from Failure: Training Debiased Classifier from Biased Classifier · Work ID: nam2020learning.
Auxiliary teachers (auxiliary_teachers); biased learner (biased_learner); Curriculum & training difficulty (curriculum_training_difficulty); failure weighting (failure_weighting); General learning & inference (general_learning_inference); Invariance & spurious factors (invariance_spurious_factors); learning difficulty (learning_difficulty); Robust learning (robust_learning); Signal–noise separation (signalnoise_separation); spurious correlation (spurious_correlation).
Genetic expert-guided learning
Guiding Deep Molecular Optimization with Genetic Exploration · Work ID: ahn2020guiding.
apprentice (apprentice); crossover (crossover); Exploration–exploitation (explorationexploitation); GEGL (gegl); Genetic algorithms (genetic_algorithms); Local search (local_search); Molecules & drug discovery (molecules_drug_discovery); mutation (mutation); penalized logP (penalized_logp); priority queue (priority_queue); Replay & off-policy reuse (replay_off_policy_reuse); Representation learning (representation_learning); Search & optimization (search_optimization); Search–learning feedback (searchlearning_feedback); Simulation / proxy-to-reality gap (simulation_proxy_to_reality_gap).
Learning what to defer
Learning What to Defer for Maximum Independent Sets · Work ID: ahn2020learning.
adaptive stages (adaptive_stages); Adaptive trajectories (adaptive_trajectories); Combinatorial optimization (combinatorial_optimization); Constraint-aware design (constraint_aware_design); deferred decisions (deferred_decisions); Graph neural networks (graph_neural_networks); Graphs & discrete problems (graphs_discrete_problems); Local search (local_search); locally decomposable objective (locally_decomposable_objective); maximum independent set (maximum_independent_set); Reinforcement learning (reinforcement_learning); Sample efficiency (sample_efficiency); Search & optimization (search_optimization).
Variational information distillation
Variational Information Distillation for Knowledge Transfer · Work ID: ahn2019variational.
CNN-to-MLP (cnn_to_mlp); Compression (compression); Compression & compact coding (compression_compact_coding); General learning & inference (general_learning_inference); heterogeneous architecture (heterogeneous_architecture); Information routing (information_routing); Knowledge distillation (knowledge_distillation); mutual information (mutual_information); Probabilistic inference (probabilistic_inference); Representation learning (representation_learning); variational bound (variational_bound).
Bucket renormalization
Bucket-Renormalization for Approximate Inference · Work ID: ahn2018bucketrenormalization.
Compression (compression); Factorization & decomposition (factorization_decomposition); General learning & inference (general_learning_inference); Learned / classical hybrids (learned_classical_hybrids); low-rank projection (low_rank_projection); mini-bucket (mini_bucket); partition function (partition_function); Probabilistic inference (probabilistic_inference); tensor network (tensor_network); Unnormalized distributions (unnormalized_distributions); variable elimination (variable_elimination).
Gauged variational inference
Gauging Variational Inference · Work ID: ahn2017gauging.
Factorization & decomposition (factorization_decomposition); Forney graph (forney_graph); gauge transformation (gauge_transformation); General learning & inference (general_learning_inference); mean field (mean_field); partition function (partition_function); Probabilistic inference (probabilistic_inference); Symmetry (symmetry); Unnormalized distributions (unnormalized_distributions).
Gauged mini-bucket elimination
Gauged Mini-Bucket Elimination for Approximate Inference · Work ID: ahn2018gauged.
Factorization & decomposition (factorization_decomposition); gauge optimization (gauge_optimization); General learning & inference (general_learning_inference); lower bound (lower_bound); Probabilistic inference (probabilistic_inference); Symmetry (symmetry); Unnormalized distributions (unnormalized_distributions); upper bound (upper_bound); weighted mini-bucket (weighted_mini_bucket).
Odd-cycle matching BP
Maximum Weight Matching using Odd-sized Cycles: Max-Product Belief Propagation and Half-Integrality · Work ID: ahn2018maximum.
Combinatorial optimization (combinatorial_optimization); Constraint-aware design (constraint_aware_design); cutting plane (cutting_plane); General learning & inference (general_learning_inference); Graph structure & transformations (graph_structure_transformations); Graphs & discrete problems (graphs_discrete_problems); half-integrality (half_integrality); LP relaxation (lp_relaxation); maximum-weight matching (maximum_weight_matching); Message passing (message_passing); odd-cycle constraint (odd_cycle_constraint); Probabilistic inference (probabilistic_inference); Search & optimization (search_optimization).
MCMC + belief propagation
Synthesis of MCMC and Belief Propagation · Work ID: ahn2016synthesis.
BP correction (bp_correction); Energy-based sampling (energy_based_sampling); General learning & inference (general_learning_inference); hard-core model (hard_core_model); Ising (ising); Learned / classical hybrids (learned_classical_hybrids); loop calculus (loop_calculus); MCMC (mcmc); Message passing (message_passing); partition function (partition_function); Probabilistic inference (probabilistic_inference); Unnormalized distributions (unnormalized_distributions).
Blossom belief propagation
Minimum Weight Perfect Matching via Blossom Belief Propagation · Work ID: ahn2015minimum.
blossom contraction (blossom_contraction); blossom expansion (blossom_expansion); Combinatorial optimization (combinatorial_optimization); Constraint-aware design (constraint_aware_design); distributed algorithm (distributed_algorithm); General learning & inference (general_learning_inference); Graph structure & transformations (graph_structure_transformations); Graphs & discrete problems (graphs_discrete_problems); linear program (linear_program); Message passing (message_passing); perfect matching (perfect_matching); Probabilistic inference (probabilistic_inference); Search & optimization (search_optimization).