Research map / Architectures and efficiency / RNNs and normalization
Neural Modulation for Flash Memory: An Unsupervised Learning Framework for Improved Reliability
Jonathan Zedaka, Elisha Halperin, Evgeny Blaichman, Amit Berman
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 |
|---|---|---|---|
| use learned neighbor-aware vth targets instead of default PAM targets | NAND flash programming · TLC, average block BER | BER reduction: 56% | Improved |
| use learned neighbor-aware vth targets instead of default PAM targets | NAND flash programming · QLC, average block BER | BER reduction: 17% | Improved |
| use learned targets instead of default targets | NAND flash programming · TLC, worst-case WL | BER reduction: 36% | Improved |
| use learned targets instead of default targets | NAND flash programming · QLC, worst-case WL | BER reduction: 9% | Improved |
| use learned targets instead of default targets | NAND flash block · PE cycles to EOL | lifetime: 25% | Improved |
| pre-distort targets with neighbor symbols | NAND flash programming · short retention time, QLC | BER | Hurt |
| pre-distort targets with neighbor symbols | NAND flash programming · long retention time, QLC | BER | Improved |
| condition cWGAN on neighbor targets | channel model · QLC | ICI modeling | Improved |
| evaluate cWGAN on shifted targets unseen in training | channel model · QLC BER vs target shift | BER prediction error: up to 5% | No clear effect |
| compare GAN predicted vs empirical BER reduction | channel model · TLC average | BER reduction: 54% vs 56% | No clear effect |
Closest papers by meaning
- Analog In-memory Training on General Non-ideal Resistive Elements: The Impact of Response Functions (NeurIPS 2025 · similarity 0.85)
- NeuralFuse: Learning to Recover the Accuracy of Access-Limited Neural Network Inference in Low-Voltage Regimes (NeurIPS 2024 · similarity 0.85)
- AtomNet: Designing Tiny Models from Operators Under Extreme MCU Constraints (AAAI 2025 · similarity 0.85)
- Sorbet: A Neuromorphic Hardware-Compatible Transformer-Based Spiking Language Model (ICML 2025 · similarity 0.84)
- S$^2$NN: Sub-bit Spiking Neural Networks (NeurIPS 2025 · similarity 0.84)
- Blending Complementary Memory Systems in Hybrid Quadratic-Linear Transformers (NeurIPS 2025 · similarity 0.84)
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.
