arXiv subjects

Xiaohan Shan

Publications and source records attributed to Xiaohan Shan.

11 recordsLinked to original sources

arXiv2026

QSimAdv: A Late-Bound, Vendor-Agnostic Architecture for High-Performance Quantum-Circuit Simulation

Portability in high-performance quantum-circuit simulation need not begin at the kernel. We present QSimAdv, which makes late binding, rather than a common kernel, the basis of vendor independence. Representation, operator lowering, and data placement are bound only when their required inputs become available. Before full-state allocation, circuit, noise, and output inspection can route eligible generic sampled-count requests to a stabiliser tableau; explicitly requested representations remain fixed. For full-state execution, backend constraints shape fusion; an ordered fused operator binds to a native lowering only after its physical targets are known. A first-class logical-to-physical layout map records non-canonical order across local and rank-address bits, so the dispatcher moves nonlocal targets only on demand. GPU, CPU, and Message Passing Interface (MPI) backends share these semantics while retaining native execution paths. We realize this design on NVIDIA GH200 and AMD MI250X/EPYC systems across local and distributed execution. With matched complex 32-bit floating-point state storage, QSimAdv leads both Aer Hopper configurations at $N=32$ and Aer's HIP backend at four shared MI250X sizes from $N=24$ to 30. Strong scaling exposes platform dependence: on setonix, QSimAdv leads both GPU and CPU comparisons at every measured rank, achieving $3.4\times$ and $2.8\times$ speedups, respectively, from one to eight ranks; neither the GH200 path nor the CPU path speeds up at eight ranks. Weak scaling reaches 256 ranks with 2 TiB GPU and 1 TiB CPU states. Together, these results support that portability can reside above the kernel boundary while execution remains native and extends across distributed memory.

quant-ph↗

arXiv2026

QAdapt: A Noise-Adaptive Neural Pre-Decoding Framework for Quantum Error Correction

Fault-tolerant quantum computing (FTQC) relies on quantum error correction to suppress physical errors and preserve logical information at scale. In practice, however, performance is constrained not only by physical noise but also by the latency of classical decoders processing rapidly generated syndrome data. This challenge is exacerbated by hardware noise that is strong, heterogeneous, and nonstationary, as well as by the simulation-to-hardware distribution shift that can substantially degrade fixed neural decoders. We present QAdapt, a noise-adaptive neural pre-decoding framework for surface-code quantum error correction. QAdapt captures local spatiotemporal correlations in syndrome data, sequentially adapts to evolving noise conditions while mitigating catastrophic forgetting, and forwards the residual syndrome to a conventional global decoder. Across 110 synthetic out-of-distribution noise configurations for rotated surface-code memory circuits, QAdapt consistently reduces the logical error rate relative to the neural pre-decoding baseline. On Google's Willow benchmark data, without target-domain fine-tuning, it achieves reductions of up to 5.79 percent in logical error rate and 9.32 percent in backend decoding latency on the residual syndrome. These results demonstrate that QAdapt provides a practical and decoder-compatible approach to improving the robustness and backend decoding efficiency of quantum error correction under evolving hardware noise.

cs.LG↗

arXiv2025

$Agent^2$: An Agent-Generates-Agent Framework for Reinforcement Learning Automation

Reinforcement learning (RL) agent development traditionally requires substantial expertise and iterative effort, often leading to high failure rates and limited accessibility. This paper introduces Agent$^2$, an LLM-driven agent-generates-agent framework for fully automated RL agent design. Agent$^2$ autonomously translates natural language task descriptions and environment code into executable RL solutions without human intervention. The framework adopts a dual-agent architecture: a Generator Agent that analyzes tasks and designs agents, and a Target Agent that is automatically generated and executed. To better support automation, RL development is decomposed into two stages, MDP modeling and algorithmic optimization, facilitating targeted and effective agent generation. Built on the Model Context Protocol, Agent$^2$ provides a unified framework for standardized agent creation across diverse environments and algorithms, incorporating adaptive training management and intelligent feedback analysis for continuous refinement. Extensive experiments on benchmarks including MuJoCo, MetaDrive, MPE, and SMAC show that Agent$^2$ outperforms manually designed baselines across all tasks, achieving up to 55\% performance improvement with consistent average gains. By enabling a closed-loop, end-to-end automation pipeline, this work advances a new paradigm in which agents can design and optimize other agents, underscoring the potential of agent-generates-agent systems for automated AI development.

cs.AI↗

arXiv2025

PhaseNAS: Language-Model Driven Architecture Search with Dynamic Phase Adaptation

Neural Architecture Search (NAS) is challenged by the trade-off between search space exploration and efficiency, especially for complex tasks. While recent LLM-based NAS methods have shown promise, they often suffer from static search strategies and ambiguous architecture representations. We propose PhaseNAS, an LLM-based NAS framework with dynamic phase transitions guided by real-time score thresholds and a structured architecture template language for consistent code generation. On the NAS-Bench-Macro benchmark, PhaseNAS consistently discovers architectures with higher accuracy and better rank. For image classification (CIFAR-10/100), PhaseNAS reduces search time by up to 86% while maintaining or improving accuracy. In object detection, it automatically produces YOLOv8 variants with higher mAP and lower resource cost. These results demonstrate that PhaseNAS enables efficient, adaptive, and generalizable NAS across diverse vision tasks.

cs.LG↗

arXiv2025

LERO: LLM-driven Evolutionary framework with Hybrid Rewards and Enhanced Observation for Multi-Agent Reinforcement Learning

Multi-agent reinforcement learning (MARL) faces two critical bottlenecks distinct from single-agent RL: credit assignment in cooperative tasks and partial observability of environmental states. We propose LERO, a framework integrating Large language models (LLMs) with evolutionary optimization to address these MARL-specific challenges. The solution centers on two LLM-generated components: a hybrid reward function that dynamically allocates individual credit through reward decomposition, and an observation enhancement function that augments partial observations with inferred environmental context. An evolutionary algorithm optimizes these components through iterative MARL training cycles, where top-performing candidates guide subsequent LLM generations. Evaluations in Multi-Agent Particle Environments (MPE) demonstrate LERO's superiority over baseline methods, with improved task performance and training efficiency.

cs.LG↗

arXiv2025

Lifelong Reinforcement Learning with Similarity-Driven Weighting by Large Models

Lifelong Reinforcement Learning (LRL) holds significant potential for addressing sequential tasks, but it still faces considerable challenges. A key difficulty lies in effectively preventing catastrophic forgetting and facilitating knowledge transfer while maintaining reliable decision-making performance across subsequent tasks in dynamic environments. To tackle this, we propose a novel framework, SDW (Similarity-Driven Weighting Framework), which leverages large-language-model-generated dynamic functions to precisely control the training process. The core of SDW lies in two functions pre-generated by large models: the task similarity function and the weight computation function. The task similarity function extracts multidimensional features from task descriptions to quantify the similarities and differences between tasks in terms of states, actions, and rewards. The weight computation function dynamically generates critical training parameters based on the similarity information, including the proportion of old task data stored in the Replay Buffer and the strategy consistency weight in the loss function, enabling an adaptive balance between learning new tasks and transferring knowledge from previous tasks. By generating function code offline prior to training, rather than relying on large-model inference during the training process, the SDW framework reduces computational overhead while maintaining efficiency in sequential task scenarios. Experimental results on Atari and MiniHack sequential tasks demonstrate that SDW significantly outperforms existing lifelong reinforcement learning methods.

cs.LG↗

arXiv2022

Higher order monotonicity and submodularity of influence in social networks: from local to global

Kempe, Kleinberg and Tardos (KKT) proposed the following conjecture about the general threshold model in social networks: local monotonicity and submodularity imply global monotonicity and submodularity. That is, if the threshold function of every node is monotone and submodular, then the spread function $σ(S)$ is monotone and submodular, where $S$ is a seed set and the spread function $σ(S)$ denotes the expected number of active nodes at termination of a diffusion process starting from $S$. The correctness of this conjecture has been proved by Mossel and Roch. In this paper, we first provide the concept AD-k (Alternating Difference-$k$) as a generalization of monotonicity and submodularity. Specifically, a set function $f$ is called \adk if all the $\ell$-th order differences of $f$ on all inputs have sign $(-1)^{\ell+1}$ for every $\ell\leq k$. Note that AD-1 corresponds to monotonicity and AD-2 corresponds to monotonicity and submodularity. We propose a refined version of KKT's conjecture: in the general threshold model, local AD-k implies global AD-k. The original KKT conjecture corresponds to the case for AD-2, and the case for AD-1 is the trivial one of local monotonicity implying global monotonicity. By utilizing continuous extensions of set functions as well as social graph constructions, we prove the correctness of our conjecture when the social graph is a directed acyclic graph (DAG). Furthermore, we affirm our conjecture on general social graphs when $k=\infty$.

cs.SI↗

arXiv2020

Discouraging Pool Block Withholding Attacks in Bitcoins

The arisen of Bitcoin has led to much enthusiasm for blockchain research and block mining, and the extensive existence of mining pools helps its participants (i.e., miners) gain reward more frequently. Recently, the mining pools are proved to be vulnerable for several possible attacks, and pool block withholding attack is one of them: one strategic pool manager sends some of her miners to other pools and these miners pretend to work on the puzzles but actually do nothing. And these miners still get reward since the pool manager can not recognize these malicious miners. In this work, we revisit the game-theoretic model for pool block withholding attacks and propose a revised approach to reallocate the reward to the miners. Fortunately, in the new model, the pool managers have strong incentive to not launch such attacks. We show that for any number of mining pools, no-pool-attacks is always a Nash equilibrium. Moreover, with only two minority mining pools participating, no-pool-attacks is actually the unique Nash equilibrium.

cs.GT↗

arXiv2018

Coreness of Cooperative Games with Truncated Submodular Profit Functions

Coreness represents solution concepts related to core in cooperative games, which captures the stability of players. Motivated by the scale effect in social networks, economics and other scenario, we study the coreness of cooperative game with truncated submodular profit functions. Specifically, the profit function $f(\cdot)$ is defined by a truncation of a submodular function $σ(\cdot)$: $f(\cdot)=σ(\cdot)$ if $σ(\cdot)\geqη$ and $f(\cdot)=0$ otherwise, where $η$ is a given threshold. In this paper, we study the core and three core-related concepts of truncated submodular profit cooperative game. We first prove that whether core is empty can be decided in polynomial time and an allocation in core also can be found in polynomial time when core is not empty. When core is empty, we show hardness results and approximation algorithms for computing other core-related concepts including relative least-core value, absolute least-core value and least average dissatisfaction value.

cs.GT↗

arXiv2017

Cumulative Activation in Social Networks

Most studies on influence maximization focus on one-shot propagation, i.e. the influence is propagated from seed users only once following a probabilistic diffusion model and users' activation are determined via single cascade. In reality it is often the case that a user needs to be cumulatively impacted by receiving enough pieces of information propagated to her before she makes the final purchase decision. In this paper we model such cumulative activation as the following process: first multiple pieces of information are propagated independently in the social network following the classical independent cascade model, then the user will be activated (and adopt the product) if the cumulative pieces of information she received reaches her cumulative activation threshold. Two optimization problems are investigated under this framework: seed minimization with cumulative activation (SM-CA), which asks how to select a seed set with minimum size such that the number of cumulatively active nodes reaches a given requirement $η$; influence maximization with cumulative activation (IM-CA), which asks how to choose a seed set with fixed budget to maximize the number of cumulatively active nodes. For SM-CA problem, we design a greedy algorithm that yields a bicriteria $O(\ln n)$-approximation when $η=n$, where $n$ is the number of nodes in the network. For both SM-CA problem with $η<n$ and IM-CA problem, we prove strong inapproximability results. Despite the hardness results, we propose two efficient heuristic algorithms for SM-CA and IM-CA respectively based on the reverse reachable set approach. Experimental results on different real-world social networks show that our algorithms significantly outperform baseline algorithms.

cs.SI↗

arXiv2015

The Least-core and Nucleolus of Path Cooperative Games

Cooperative games provide an appropriate framework for fair and stable profit distribution in multiagent systems. In this paper, we study the algorithmic issues on path cooperative games that arise from the situations where some commodity flows through a network. In these games, a coalition of edges or vertices is successful if it enables a path from the source to the sink in the network, and lose otherwise. Based on dual theory of linear programming and the relationship with flow games, we provide the characterizations on the CS-core, least-core and nucleolus of path cooperative games. Furthermore, we show that the least-core and nucleolus are polynomially solvable for path cooperative games defined on both directed and undirected network.

cs.GT↗