arXiv subjects

Zijian Ye

Publications and source records attributed to Zijian Ye.

3 recordsLinked to original sources

arXiv2026

Parameter Efficient Machine Unlearning on Hybrid Resistive Memory based Compute-in-Memory Accelerators

Resistive memory compute-in-memory accelerators provide energy efficient analogue matrix vector multiplication for neural network inference, but frequent reprogramming of analogue weights remains costly because of device variability and iterative write and verify operations. This limitation hinders their use in edge model adaptation, including approximate machine unlearning and continual learning, where model parameters may need to be updated repeatedly in response to data deletion requests or newly arriving tasks. Here we present a co-design approach across hardware and software that maps frozen pretrained weights to analogue resistive memory arrays while placing trainable low rank adaptation branches in SRAM connected digital compute. By using LoRA style parameter efficient updates, the proposed scheme confines adaptation to a small set of digital parameters and avoids repeated reprogramming of the analogue backbone. To our knowledge, this work provides the first experimental demonstration of approximate machine unlearning on a fabricated resistive memory CIM accelerator. We validate the framework on a 180 nm 128x128 1T1R resistive-memory macro for face recognition, and through circuit-accurate simulations for speaker authentication and stylized image generation tasks, owing to the substantial model sizes involved. Compared with a baseline that directly updates analog weights, our hybrid mapping reduces analog training/update cost by up to 148x, on-chip deployment overhead by up to 388x, and inference energy by up to 59x, while preserving competitive task performance. These results show that hybrid analogue-digital LoRA mapping can enable efficient post-deployment adaptation on RM-CIM hardware, although formal machine-unlearning guarantees and large-scale system integration remain open challenges.

cs.ET↗

arXiv2026

Analysis of Nuclear Fragmentation Functions for Pions with $A$ and $ν$ Dependence

We present a QCD analysis of pion nuclear fragmentation functions (nFFs), which encode nuclear modifications to hadronization in high-energy nuclear collisions. Within this framework, vacuum fragmentation functions and their nuclear modifications are extracted simultaneously. The nuclear effects are parameterized as functions of the mass number $A$, the energy of the fragmenting parton in the target rest frame $ν$, and the hadron energy fraction $z$, allowing their dependence on these variables to be quantified. Our analysis includes semi-inclusive deep-inelastic scattering data on nuclear targets, with kinematic cuts chosen to ensure the applicability of perturbative QCD and collinear factorization. The resulting fit provides a good description of most datasets, with the nFFs well constrained in the energy fraction range $z \in [0.2, 0.7]$. Additionally, with our new nFFs, we present next-to-leading order predictions in $pp$ and $pA$ collisions, which show reasonable agreement with ALICE data within the current experimental uncertainties.

hep-ph↗

arXiv2026

DBellQuant: Breaking the Bell with Double-Bell Transformation for LLMs Post Training Binarization

Large language models (LLMs) demonstrate remarkable performance but face substantial computational and memory challenges that limit their practical deployment. Quantization has emerged as a promising solution; however, its effectiveness is often limited by quantization errors arising from weight distributions that are not quantization-friendly and the presence of activation outliers. To address these challenges, we introduce DBellQuant, an innovative post-training quantization (PTQ) framework that achieves nearly 1-bit weight compression and 6-bit activation quantization with minimal performance degradation. DBellQuant uses Learnable Transformation for Dual-Bell (LTDB) algorithm, which transforms single-bell weight distributions into dual-bell forms to reduce binarization errors and applies inverse transformations to smooth activations. DBellQuant sets a new state-of-the-art by preserving superior model performance under aggressive weight and activation quantization. For example, on the Wikitext2 dataset, DBellQuant achieves a perplexity of 14.39 on LLaMA2-13B with 6-bit activation quantization, significantly outperforming BiLLM's 21.35 without activation quantization, underscoring its potential in compressing LLMs for real-world applications.

cs.LG↗