The International Conference on Learning Representations. Explore 223 oral papers as a constellation of knowledge — hover to discover, click to dive deep.
Each dot is an oral paper. Clusters are research territories. Hover to peek, click to explore. Gold stars are award winners, blue rings are papers we’ve turned into interactive lessons.
Two outstanding papers and one honorable mention, selected from 4,068 submissions.
A tiny transformer can encode patterns that would require an exponentially larger RNN — and verifying even simple properties about it is provably intractable (EXPSPACE-complete).
◆ Read our interactive lesson →39% average performance degradation across 15 LLMs when tasks are delivered gradually across turns. Even reasoning models like o3 and DeepSeek-R1 suffer.
◆ Read our interactive lesson →Optimal polynomial approximations for polar decomposition — provably best convergence for the Muon optimizer via approximation theory and the Equioscillation Theorem.
◆ Read our interactive lesson →We’ve processed these ICLR 2026 papers into interactive Veanors lessons — full derivations, canvas simulations, and quizzes.
Exponential succinctness gaps between transformers and RNNs. Verification is EXPSPACE-complete.
Sharded simulation reveals 39% degradation when tasks spread across turns.
Optimal polynomial approximations for polar decomposition in the Muon optimizer.
Inference-first SSM with exp-trapezoidal discretization, complex states, and MIMO updates.
Contrastive loss disentangles semantic from syntactic features in a self-supervised manner.
First benchmark for world models in closed-loop agent-environment interactions.
Multi-view depth and ray prediction from any number of input views with a plain ViT.
DPO as weighted KL-projection causes preference reversal. AuxDPO fixes it.
Predict via gradient-descent energy minimization. 35% faster scaling than Transformer++.
40 workshops, 2,863 accepted papers. Each cluster is a workshop — hover to see papers, click to open on ICLR or OpenReview.
All papers organized by research territory. Expand a cluster to browse.