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
155 results for “keyword:"experiment"” in ICLR 2027. 30 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Invent the Questions, Compile the Consequences: Learning Executable Structure from Terminal Feedback▲9.2k2.67
[PDF]Seeds Do Not Replicate Runs: The Unit of Replication, the Estimand and the Budget Rule in Agent RL▼4.5k2.67
[PDF]Beyond Terminal Objectives: Bayesian Experimental Design with Multistep Actions and Losses▲3972.67
[PDF]MatGym: An Agent-Centric Environment Bridging Materials Discovery and Reproducible Synthesis▼4.5k2.67
[PDF]What Success Rate Hides: Mapping Disturbance Tolerance in Vision-Language-Action Policies▼4.3k2.00
[PDF]Beyond Fixed Acquisition Functions: Temporal Difference Active Learning for Perturbation Discovery▼4.3k2.00
[PDF]Learning Treatment Effects during Resource Allocation via Priority-Queue Randomization▼4.3k2.00
[PDF]TerraWatt: Discovering Hidden World Dynamics Through Irreversible Capital Allocation in a Physically Grounded Voxel Environment▼4.3k2.00
[PDF]Can domain synergy mislead? Measurement confounds in multi-domain RLVR interaction estimation▼1.2k1.67
[PDF]What Can Large Language Models Discover from Experiments? Scientific Claims and Discovery Value across Multiple Worlds▼1611.50
[PDF]Marginal normality does not constrain time: a prescribed Gaussian path law for JEPA world models▼1361.33
[PDF]Which Experiment Would Change Your Mind? Bayesian Optimal Experimental Design for Connectome-Constrained Models▼1.2k1.67
Withdrawn or desk rejected (30)
[PDF]Caustic Spectroscopy: Identifiability and Acquisition Limits of Topological Event TimingWithdrawn2.00
[PDF]Independence over Accuracy: Generator–Verifier Error Correlation in Recursive Self-Improvement in Small Language ModelsWithdrawn1.67
[PDF]Ace Olympiad Math, Fail Simple Arithmetic: Systematic Capability Unreliability in LLMsDesk rejected
[PDF]What Agent Memory Is Worth: A Zero-Cost Randomised Experiment Inside Test-Time ScalingDesk rejected