ICLR 2025PastAI for science
ICLR 2025 Workshop: XAI4Science: From Understanding Model Behavior to Discovering New Scientific Knowledge
ICLR 2025 Workshop XAI4Science
- Submission deadline
- Feb 11, 2025, 11:59 UTCimported from OpenReview — check the website for extensions
- Submission portal
- OpenReview
- Notes
- Topics were auto-suggested and may be imprecise — edits welcome.
Accepted papers (40)
Fetched from OpenReview (v2) on 2026-06-10.
$\text{CO}_2$-Net: A Physics-Informed Spatio-Temporal Model for Global $\text{CO}_2$ Reconstruction
AlphaGo or beta-hCG: a reinforcement learning framework for assisted conception
Automated Capability Discovery via Model Self-Exploration
BarkXAI: A Lightweight Post-Hoc Explainable Method for Tree Species Classification with Quantifiable Concepts
Bayesian Concept Bottleneck Models with LLM Priors
Causal Concept Graph Models: Beyond Causal Opacity in Deep Learning
Causal Lifting of Neural Representations: Zero-Shot Generalization for Causal Inferences
Causally Reliable Concept Bottleneck Models
Circuit mechanism for compositional induction in transformer
Counterfactual Concept Bottleneck Models
Efficient and Flexible Neural Network Training through Layer-wise Feedback Propagation
Emergence of Computational Structure in a Neural Network Physics Simulator
From Markov to Laplace: How Mamba In-Context Learns Markov Chains
Generating $\pi$-Functional Molecules Using STGG+ with Active Learning
Graph Discrete Diffusion: a Spectral Study
Hybrid Generative Modeling for Incomplete Physics: Deep Grey-Box Meets Optimal Transport
LEARNING MULTIPHASE AND MULTIPHYSICS SYSTEM WITH DECOUPLED STATE SPACE MODEL
LENS: Learning and Evolving Numerical Scores for Cohort-Specific Clinical Insights
Limits of Deep Learning: Sequence Modeling through the Lens of Complexity Theory
Machine learning-based Optimization for Molten pool Dynamics in Laser Manufacturing
Massive Activations in Graph Neural Networks: Decoding Attention for Domain-Dependent Interpretability
Measuring Leakage in Concept-Based Methods: An Information Theoretic Approach
Modeling Multi-Regional and Non-Stationary Neural Dynamics via Latent Sub-Circuits
Moment Neural Operator: Interpretable mapping in discontinuous function spaces
NeuralDEM: Real-time Simulation of Industrial Particulate Flows
Piecewise Polynomial Regression of Tame Functions via Integer Programming
Post-hoc Interpretability Illumination for Scientific Interaction Discovery
Reconstructing Dynamics from Steady Spatial Patterns with Partial Observations
Rethinking Visual Counterfactual Explanations Through Region Constraint
SAeUron: Interpretable Concept Unlearning in Diffusion Models with Sparse Autoencoders
Scaling Sparse Autoencoders for Interpreting Protein Structure Prediction
SHAP-BASED A-POSTERIORI INTERPRETABILITY FOR GRAPH NEURAL NETWORKS IN CFD-BASED SUSTAINABLE BUILDING SIMULATIONS
Skip the Equations: Learning Behavior of Personalized Dynamical Systems Directly From Data
Spatially-Informed Sampling Enables Accurate Prediction of Large-Scale Mutational Effects
TIME-AWARE FEATURE SELECTION: ADAPTIVE TEMPORAL MASKING FOR STABLE SPARSE AUTOENCODER TRAINING
TIMING: Temporality-Aware Integrated Gradients for Time Series Explanation
Towards Mechanistic Interpretability of Graph Transformers via Attention Graphs
ULTra: Unveiling Latent Token Interpretability in Transformer-Based Understanding
Unveiling the Hidden Structure of Self-Attention via Kernel Principal Component Analysis
Why Uncertainty Calibration Matters for Reliable Perturbation-based Explanations