Diffusion language models generate multiple token positions in parallel, but few-step decoding makes parallel commitment difficult: tokens that appear reliable individually may be jointly inconsistent, while premature commitments can prevent later context-dependent refinement. In this paper, we introduce SetCommit, a t...
Yuan-Nuo Feng, Yu-Bing Ye, Wen-Shuai Yao et al.· 0 citations
Real-time super-resolution (SR) remains challenging for edge devices because deep-learning-based methods require substantial multiply-accumulate (MAC) operations, resources, and power. Lookup-table (LUT)-based SR reduces computation by replacing convolutional inference with table queries, but existing methods still suf...
Bo-Yu Li, Chen-Chen Ding, Zhi-Lin Ai et al.· 0 citations
Transformer-based models have become the backbone of a wide range of tasks, but their intensive computation and data movement incur substantial energy overhead. Analog compute-in-memory (CIM) offers an energy-efficient alternative, yet its inherent hardware non-idealities perturb stored weights and can significantly de...
Wen-Shuai Yao, Wen-Yong Zhou, Han-Yong Shao et al.· 0 citations
NoVA-CIM is presented, a noise- and correlation-tolerant stochastic interface for analog CIM by replacing multi-bit ADC readout with random-reference 1-bit sensing and reconstructing results through lightweight counting, which provides a unified robustness analysis showing that dynamic read noise is suppressed through...
Jia-Chen Ren, Wen-Shuai Yao, Hao-Bo Liu et al.· 0 citations
This work investigates diffusion inference using a noise model calibrated and validated against measurements collected from multiple physical CIM chips, and proposes ASSERT, a training-free sampler that uses higher stochasticity early and smoothly transitions to deterministic denoising.
Yuan-Nuo Feng, Yi-Zhe Chen, Wen-Shuai Yao et al.· 0 citations
A hierarchical token protection strategy is proposed that keeps sink tokens and a sliding recent-token window on a higher-precision digital path while processing the bulk KV cache on analog CIM, revealing that initial and recent tokens exhibit disproportionate vulnerability to hardware noise.
Yuan-Nuo Feng, Wen-Yong Zhou, Yuang Ma et al.· arXiv.org· 0 citations
ReTopK is a training-free method that accelerates dynamic Top-$K$ attention by reusing historical retrieval decisions and retains the complete KV cache and reuses only selected indices, rather than historical scores, attention weights, or outputs.
The impact of analog CIM nonidealities on DiT sampling is characterized and a retraining-free, sampler-side recalibration that adjusts only the CFG scale for a given CIM condition is proposed, showing that the optimal guidance scale increases with CIM noise.
Long-term exposure to PM2.5 and O3 is associated with increased risk of ARIs, with similar associations observed for ILI, and the importance of long-term air pollution control and targeted interventions for susceptible populations, particularly during cold seasons is highlighted.
Xiaole Duan, Xiao Yu, Hongyu Liang et al.· JMIR Public Health and Surve...· 0 citations
Approximate Speculative Decoding (ASD) is introduced, a training-free verifier that replaces binary first-mismatch truncation with budgeted longest-prefix selection and reuses the contiguous target-greedy suffix without additional approximate decisions or target-model forward passes.
Yuan-Nuo Feng, Zegang Peng, Yu-Xin Xie et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
