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
842 results for “keyword:"compression"” in ICLR 2027. 183 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]What Survives Is Not Enough: Loss-Aligned Directional Salience Calibration for LLM Compression▲9.2k2.67
[PDF]Lost in Pruning: Silent Multilingual Collapse and Training-Free Recovery for One-Shot MoE Pruning▼1.2k3.00
[PDF]One Model, Multiple Compute Budgets: Multi-Ratio ConceptMoE for Flexible Compute Across Turns▼4.5k2.67
[PDF]Not All Calibration Data Are Equal: Optimizing Calibration Data for Post-Training Quantization▲3.0k2.33
[PDF]CaliSplat: Sparse Gaussian Fitting and Calibrated Rate Allocation for Image Compression▲5103.00
[PDF]TFPrune: Pruning Visual Redundancy while Preserving Token Functions for Efficient VLMs▼4.3k2.00
[PDF]When to Evict, Not What to Keep: Draft-Guided Eviction for Training-Free KV-Cache Compression▼1.2k3.00
[PDF]RACE-HE: Rotation-Aware Compression for Efficient Private Inference with Homomorphic Encryption▼5.8k2.33
[PDF]Same Score, Different State: Post-Training Moves Pre-Answer Recoverability under KV Compression▼4.3k2.00
[PDF]RoQAT: A Robust Quantization-aware Training Method for LLMs Extremely Low-bit Downstream Reasoning▼1.2k1.67
[PDF]Relative Kinetic Utility: Calibrating Cross-Layer Credit for Global Structured LLM Pruning▼5.8k2.33
[PDF]-Omni: Reducing Structural and Precision Redundancy in Omni-LLMs for Long-Form Generation▲5.4k3.00
[PDF]Communication-Efficient Agnostic Federated Learning via Faster Convergence and Compression▲5103.00
[PDF]WorldRetain: Predictive Retention and Spatial Compression for Efficient and Consistent Video World Model Inference▼4.5k2.67
[PDF]CurveTQ: Rotation-Free Trellis Quantization of LLM Weights via Curvature-Weighted Search▼4.5k2.67
[PDF]Beyond the Canvas: Efficient dLLMs via Self-Guided CoT Compression and Suffix Sparsification▼5.8k2.33
[PDF]CostBit: Candidate-Cost-Informed Bit Allocation for Post-Training Quantization of LLMs▼5.8k2.33
Withdrawn or desk rejected (183)
[PDF]Rethinking Video Token Compression with a Global Codebook: Learning Once, Compressing EverywhereWithdrawn1.67
[PDF]Preserving Decision-Relevant History: Response-Rank Memory for Write-Before-Query PredictionWithdrawn1.00