arXiv subjects

Zisheng Liu

Publications and source records attributed to Zisheng Liu.

4 recordsLinked to original sources

arXiv2026

Discrete Diffusion for Complex and Congested Multi-Agent Path Finding with Sparse Social Attention

Multi-Agent Path Finding (MAPF) is a coordination problem that requires computing globally consistent, collision-free trajectories from individual start positions to assigned goal positions under combinatorial planning complexity. In dense environments, suboptimal initial plans induce compound conflicts that hinder feasible repair. For repair-based solvers like LNS2, initial plan quality critically affects downstream repair, yet this factor remains underexplored. We propose DiffLNS, a hybrid framework that integrates a discrete denoising diffusion probabilistic model (D3PM) with LNS2. The D3PM serves as an initializer with sparse social attention that learns a spatiotemporal prior over coordinated multi-agent action trajectories from expert demonstrations and samples multiple joint plans. Operating directly on the categorical action space, our discrete diffusion preserves the MAPF action structure and samples from a multimodal joint-plan distribution to produce diverse drafts well suited for neighborhood repair. These drafts act as warm starts for downstream repair, which completes unfinished trajectories and resolves remaining conflicts under hard MAPF constraints. Experimental results show that despite being trained only on instances with at most 96 agents, the initializer generalizes to scenarios with up to 312 agents at inference time. Across 20 complex and congested settings, DiffLNS achieves an average success rate of 95.8%, outperforming the strongest tested baseline by 9.6 percentage points and matching or exceeding all baselines in all 20 settings. To the best of our knowledge, this is the first work to leverage discrete diffusion for warm-starting an LNS-based MAPF solver.

cs.AI↗

arXiv2026

Hardwired-Neurons Language Processing Units as General-Purpose Cognitive Substrates

The rapid advancement of Large Language Models (LLMs) has established language as a core general-purpose cognitive substrate, driving the demand for specialized Language Processing Units (LPUs) tailored for LLM inference. To overcome the growing energy consumption of LLM inference systems, this paper proposes a Hardwired-Neurons Language Processing Unit (HNLPU), which physically hardwires LLM weight parameters into the computational fabric, achieving several orders of magnitude computational efficiency improvement by extreme specialization. However, a significant challenge still lies in the scale of modern LLMs. A straightforward hardwiring of gpt-oss 120 B would require fabricating photomask sets valued at over 6 billion dollars, rendering this straightforward solution economically impractical. Addressing this challenge, we propose the novel Metal-Embedding methodology. Instead of embedding weights in a 2D grid of silicon device cells, Metal-Embedding embeds weight parameters into the 3D topology of metal wires. This brings two benefits: (1) a 15x increase in density, and (2) 60 out of 70 photomask layers are homogeneous across chips, including all EUV photomasks. In total, Metal-Embedding reduced the photomask cost by 112x, bringing the Non-Recurring Engineering (NRE) cost of HNLPU into an economically viable range. Experimental results show that HNLPU achieved 249,960 tokens/s (5,555x/85x that of GPU/WSE), 36 tokens/J (1,047x/283x that of GPU/WSE), 13,232 mm2 total die area, $59.46 M-123.5 M estimated NRE at 5 nm technology. Analysis shows that HNLPU achieved 41.7-80.4x improvement in cost-effectiveness and 357x reduction in carbon footprint compared to OpenAI-scale H100 clusters, under an annual weight updating assumption.

cs.AR↗

arXiv2023

Error estimation for the non-convex cosparse optimization problem

When the signal does not have a sparse structure but has sparsity under a certain transformation domain, Nam et al. \cite{NS} introduced the cosparse analysis model, which provides a dual perspective on the sparse representation model. This paper mainly discusses the error estimation of non-convex $\ell_p(0<p<1)$ relaxation cosparse optimization model with noise condition. Compared with the existing literature, under the same conditions, the value range of the $Ω$-RIP constant $δ_{7s}$ given in this paper is wider. When $p=0.5$ and $δ_{7s}=0.5$, the error constants $C_0$ and $C_1$ in this paper are better than those corresponding results in the literature \cite{Cand,LiSong1}. Moreover, when $0<p<1$, the error results of the non-convex relaxation method are significantly smaller than those of the convex relaxation method. The experimental results verify the correctness of the theoretical analysis and illustrate that the $\ell_p(0<p<1)$ method can provide robust reconstruction for cosparse optimization problems.

math.OC↗

arXiv2023

Convergence analysis on the alternating direction method of multipliers for the cosparse optimization problem

From a dual perspective of the sparse representation model, Nam et al. proposed the cosparse analysis model. In this paper, we aim to investigate the convergence of the alternating direction method of multipliers (ADMM) for the cosparse optimization problem. First, we examine the variational inequality representation of the cosparse optimization problem by introducing auxiliary variables. Second, ADMM is used to solve cosparse optimization problem. Finally, by utilizing a tight frame with a uniform row norm and building upon lemmas and the strict contraction theorem, we establish a worst-case $\mathcal{O}(1/t)$ convergence rate in the ergodic sense.

math.OC↗