Research map / Architectures and efficiency / RNNs and normalization
Efficient Mismatch-Tolerant Coding for Model-Driven Compression
Aviv Adler, Jennifer Tang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace uniform bins and loose buffers (PMATIC V1) with tight bound, non-uniform bins and message-optimized parameters (V2) | PMATIC · Emma, eps=0.02 | compression ratio: 0.2113 -> 0.0892 | Improved |
| replace V1 with V2 on Qwen 2.5 Wikipedia | PMATIC · Wikipedia, eps=0.02 | compression ratio: 0.2297 -> 0.1349 | Improved |
| replace V1 with V2 on Mistral Hamlet | PMATIC · Hamlet, eps=0.02 | compression ratio: 0.3447 -> 0.2269 | Improved |
| increase robustness eps from 0.02 to 1.0 (ablation) | PMATIC with Llama 3.1 · Emma | compression ratio: 0.0893 -> 0.2932 | Hurt |
| add PMATIC robustness (eps=0.02) over plain arithmetic coding | Llama 3.1 compression · Wikipedia | compression ratio: 0.0815 -> 0.1349 | Hurt |
| use PMATIC instead of arithmetic coding under real cross-machine non-determinism | LLM compression · Real mismatch decoding (M4 vs M2) | correctly decoded files | Improved |
| raise eps from 0.0001 to 0.003 (ablation) | PMATIC with Llama 3.1 · Emma token error rate | token error rate: 0.007758 -> 0.000610 | Improved |
| add synthetic uniform noise within [-eps, eps] to decoder logits (ablation) | PMATIC · all files, same device | decoding success | No clear effect |
Closest papers by meaning
- Synchronizing Probabilities in Model-Driven Lossless Compression (ICLR 2026 · similarity 0.95)
- L3TC: Leveraging RWKV for Learned Lossless Low-Complexity Text Compression (AAAI 2025 · similarity 0.83)
- Minimum Entropy Coupling with Bottleneck (NeurIPS 2024 · similarity 0.83)
- SpQR: A Sparse-Quantized Representation for Near-Lossless LLM Weight Compression (ICLR 2024 · similarity 0.82)
- Extreme Compression of Large Language Models via Additive Quantization (ICML 2024 · similarity 0.82)
- LittleBit: Ultra Low-Bit Quantization via Latent Factorization (NeurIPS 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
