Selected work

Representative Papers

Querying Inconsistent Prioritized Data with ORBITS: Algorithms, Implementation, and Experiments (Extended Abstract)

Feb 16, 2022Description Logics

This paper addresses inconsistent-tolerant query answering over prioritized knowledge bases—comprising logical theories, factual databases, and priority relations among conflicting facts. We systematically support query evaluation under three classical semantics—AR (cautious), IAR (intersection of all repairs), and brave—over two classes of optimal repair models: Pareto-optimal and completion-based repairs. Our key contribution is the first unified SAT encoding framework capable of handling arbitrary priority relations, enabling joint modeling and efficient solving for both repair classes and all three semantics. Based on this encoding, we implement ORBITS, a novel reasoning system. Experimental results demonstrate that ORBITS significantly outperforms baseline approaches across all semantics, highlighting the critical impact of semantic choice and solving strategy on performance. The work establishes a new paradigm for practical reasoning over inconsistent prioritized knowledge bases.

18 citationsRead paper

Going Deep and Going Wide: Counting Logic and Homomorphism Indistinguishability over Graphs of Bounded Treedepth and Treewidth

Aug 11, 2023Annual Conference for Computer Science Logic

This paper investigates the expressive power of the fragment $C^k_q$ of first-order logic with counting quantifiers—restricted to at most $k$ variables and quantifier rank at most $q$—focusing on its logical equivalence in relation to graph structure. We characterize $C^k_q$-equivalence via homomorphism indistinguishability and define the associated graph class $T^k_q$, providing its first purely graph-theoretic characterization: $T^k_q = {G mid ext{hom}(G, -) ext{ is invariant over } C^k_q ext{-equivalent graphs}}$. Our contributions are threefold: (1) an elementary, Dvořák-inspired construction reproves that two graphs are $C^k_q$-equivalent iff they are homomorphism-indistinguishable over $T^k_q$; (2) we establish that the class of graphs with treedepth at most $q$ is closed under homomorphism distinguishability; and (3) we strictly separate $T^k_q$ from $TW_{k-1} cap TD_q$ for $q gg k$, thereby refuting Roberson’s (2022) conjecture that $TD_q$ is homomorphism-distinguishability closed.

9 citations2 influentialRead paper

When LLMs Play the Telephone Game: Cumulative Changes and Attractors in Iterated Cultural Transmissions

Jul 05, 2024arXiv.org

Large language models (LLMs) exhibit systematic distortion and stabilization in information transmission across multi-round chained interactions, yet the underlying dynamics remain poorly understood. Method: Inspired by cultural evolution theory, we design a “transmission chain” experiment—mimicking a telephone game—where LLMs iteratively paraphrase texts. We quantify intergenerational evolution across dimensions including toxicity, sentiment, complexity, and length, conducting cross-model (Llama/GPT/Claude), cross-scale, and multi-prompt comparative analyses. Contribution/Results: We demonstrate that LLM populations spontaneously converge toward semantic attractors during chained interaction. Attractor strength is significantly modulated by instruction openness, model parameter count, and initial text properties. Toxicity exhibits the strongest and most stable evolutionary dominance, while sensitivity to attraction varies hierarchically across attributes. This work provides the first empirical evidence of cumulative bias and cultural attractor mechanisms in LLM collective behavior, establishing a novel paradigm for studying information propagation dynamics among foundation models.

6 citations1 influentialRead paper

Impact of knowledge on the cost of treasure hunt in trees

Aug 07, 2021Networks

This paper studies the edge-traversal cost of treasure-hunt tasks by mobile agents on tree networks, focusing on how initial knowledge—specifically, map availability (complete vs. blind) and target-distance awareness (known vs. unknown)—affects worst-case efficiency. Methodologically, it establishes, for the first time, a rigorous partial order among the four knowledge classes; combines deterministic distributed algorithms, graph-traversal analysis, and competitive ratio theory; and derives tight upper and lower bounds on the cost penalty induced by decreasing knowledge precision. Key results show that when distance is known, blind-map traversal incurs significantly higher cost than complete-map traversal; when distance is unknown, the gap narrows substantially; and with a complete map but unknown distance, the cost increase is intermediate. All bounds are tight and constructively achievable. The findings reveal an intrinsic asymmetry in how distance information and map granularity affect efficiency, providing a theoretical benchmark for knowledge–performance trade-offs in distributed search.

3 citationsRead paper

Promises, Perils, and (Timely) Heuristics for Mining Coding Agent Activity

Jan 26, 2026

This study addresses the lack of systematic empirical evidence regarding the real-world impact of coding agents in software development. Leveraging Mining Software Repositories (MSR) methods, it presents the first large-scale analysis of activity traces from large language model–based coding agents on GitHub, systematically identifying their behavioral patterns, potential risks, and effective usage strategies in authentic development environments. The research yields a set of empirically grounded insights concerning optimal timing for agent adoption, reliability concerns, and practical heuristics for deployment. These findings fill a critical gap in the literature, offering actionable guidance for developers and establishing a foundation for future investigations into AI-assisted programming.

2 citationsRead paper

Recent publications

Latest Papers

AUPE: Collaborative byzantine fault-tolerant peer-sampling

Sep 16, 2026

本文提出AUPE协议，利用可信节点协作追踪系统中标识符的传播，有效解决拜占庭节点在视图中的过度表示问题，提高分布式系统的容错能力。

0 citationsRead paper

From powder to part: influence of virgin and recovered Inconel 625 powders on the DED-LP processability, microstructure and mechanical properties

Sep 16, 2026

研究通过分析Inconel 625粉末的重用对激光粉末直接能量沉积工艺、微观结构及力学性能的影响，揭示了热等静压处理改善延展性但降低拉伸强度，并指出需全面评估粉末重用。

0 citationsRead paper

Interpretable Multi-Instance Learning Enables Early Prediction of Key Molecular Alterations from Routine Flow Cytometry in Acute Myeloid Leukemia

Sep 16, 2026

研究使用基于决策树的多实例学习分类器，通过常规流式细胞术数据预测急性髓系白血病中的NPM1和FLT3-ITD突变状态，实现早期诊断。

0 citationsRead paper

Bridging the Gap Between Plain VASS and Branching VASS

Sep 14, 2026

研究解决了BVASS的可达性问题，通过引入新的wqo和利用合并属性，证明了BVASS的可达集与VASS相似，从而推导出关于BVASS可达集几何结构的几个结果。

0 citationsRead paper

Ramsey Obstructions to Disambiguation

Sep 14, 2026

本文通过构造特定部分矩阵，解决线性分类器的消歧问题，并利用Ramsey理论证明了无论怎样消歧都会包含所有指定大小的二进制矩阵。

0 citationsRead paper