NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of ICLR 2027
15,063 results for “keyword:"ode"” in ICLR 2027. 3,642 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]MPMWorlds: A Dataset of Executable Material-Point-Method Simulations for Inferring and Extrapolating Physical Dynamics▼4.5k2.67
[PDF]Multi-Level Transformer: A Minimal Transformer Extension for Learned Sequence Chunking▲9.2k2.67
[PDF]Beckmann World Models: Direct Terminal-Map Learning for One-Step Action-Conditioned Prediction▼4.3k2.00
[PDF]IDEAL-KV: Inference-Aligned Codebook Learning and Calibrated Stage Allocation for KV-Cache Compression▲5.4k3.00
[PDF]Group-Invariant Statistics Determine Embedding Geometry: Harmonic Analysis of Representations from Bach to the Night Sky▲3892.67
[PDF]When Sharpening Becomes Collapse: Sampling Bias and Semantic Coupling in RL with Verifiable Rewards▼1.2k1.67
[PDF]IsoAlign: Isoform-Aware DNA-Protein Alignment via Joint-Embedding Predictive Architectures▼4.3k2.00
[PDF]RELACE: RETROSPECTIVE LIKELIHOOD-BASED ACTION CREDIT ESTIMATION FOR LONG-HORIZON LANGUAGE AGENTS▼1.2k1.67
[PDF]HIDDENFACE: STREAMING CONVERSATIONAL FACIAL MOTION FROM HIDDEN STATES OF A LARGE AUDIOLANGUAGE MODEL▲3892.67
[PDF]Opt-Judger: Fine-Grained Evaluation of LLM-Based Optimization Modeling via Feasibility Verification▼5.9k2.33
[PDF]Diagnosing & Mitigating Constraint Induced Response Collapse in Instruction-Tuned LLMs▼4.5k2.67
[PDF]Where Should a Fixed Budget of Human Data Go? Allocating Across Time and Modes in Recursive Training▲3.9k1.67
Withdrawn or desk rejected (3,642)
[PDF]Mind Your Marginals: What Governs Fidelity in Linearized Hierarchical Optimal Transport?Desk rejected1.00
[PDF]A Window Is Not a Door: Mapping the Limits and Conditions of Knowledge Editing in Language ModelsDesk rejected2.33
[PDF]Propagation Is Not Decoration: When Physical Diffraction Carries Generative ComputationWithdrawn1.00