arXiv subjects

Piaohong Wang

Publications and source records attributed to Piaohong Wang.

16 recordsLinked to original sources

arXiv2026

PERFOPT-Bench: Evaluating Coding Agents on Software Performance Optimization

Coding-agent benchmarks have largely measured whether agents can produce functionally correct patches, but production software also demands measurable speedups on real execution targets. Performance optimization is a distinct agentic task: agents must profile executions, diagnose cross-layer bottlenecks, edit code without breaking correctness, and verify that gains are reproducible rather than measurement artifacts. We introduce PERFOPT-Bench, a benchmark for evaluating this full performance-engineering loop. Each task provides a correct but deliberately suboptimal codebase and asks the agent to improve a target performance metric; scoring requires hidden correctness tests, verified-speedup measurement, and trajectory-level audit. We evaluate 7 agent stacks with different LLMs and agent frameworks on 7 long-horizon optimization tasks. The results show that optimization performance is workload-dependent rather than determined by model identity alone: no single stack dominates, and changing the agent framework can materially change the same LLM's per-task speedup profile. We further find that raw speedup is unsafe as a benchmark score, since some large gains arise from benchmark-specific shortcut exploitation; an exploratory relay pilot suggests that restarting from an externalized optimization summary can recover additional headroom after an initial session stops. The benchmark and our evaluation are available at: https://anonymous.4open.science/r/Dataset-D3CC.

cs.SE↗

arXiv2026

Mixed-Initiative Context: Structuring and Managing Context for Human-AI Collaboration

In the human-AI collaboration area, the context formed naturally through multi-turn interactions is typically flattened into a chronological sequence and treated as a fixed whole in subsequent reasoning, with no mechanism for dynamic organization and management along the collaboration workflow. Yet these contexts differ substantially in lifecycle, structural hierarchy, and relevance. For instance, temporary or abandoned exchanges and parallel topic threads persist in the limited context window, causing interference and even conflict. Meanwhile, users are largely limited to influencing context indirectly through input modifications (e.g., corrections, references, or ignoring), leaving their control neither explicit nor verifiable. To address this, we propose Mixed-Initiative Context, which reconceptualizes the context formed across multi-turn interactions as an explicit, structured, and manipulable interactive object. Under this concept, the structure, scope, and content of context can be dynamically organized and adjusted according to task needs, enabling both humans and AI to actively participate in context construction and regulation. To explore this concept, we implement Contextify as a probe system and conduct a user study examining users' context management behaviors, attitudes toward AI initiative, and overall collaboration experience. We conclude by discussing the implications of this concept for the HCI community.

cs.HC↗

arXiv2026

"Law at Your Fingertips": Understanding Legal Information Seeking on Video-Sharing Platforms in China

Equipping laypeople with the capabilities to seek legal information has been an important goal for Legal Empowerment in modern society. However, unlike general information-seeking behaviors, legal information seeking is characterized by high stakes, urgency, and a critical need for emotional support, which traditional text-based searching platforms struggle to satisfy. In recent years, people have been increasingly turning to Video-Sharing Platforms (VSPs) for access to legal information and to fulfill their legal needs. Despite the importance of this shift, such VSP-mediated legal information-seeking practices remain underexplored. Through an observational analysis of legal content on two VSPs (Douyin and Bilibili) and interviews with 20 Chinese information seekers, this study examined the practices and challenges associated with seeking, comprehending, and evaluating legal information on VSPs. We further revealed the formation of trust and engagement on the VSP-based legal knowledge-sharing community, highlighting how VSP affordances helped mitigate seekers' epistemic discomfort and satisfy their needs for emotional support. In the discussion, we provided insights on balancing heuristic and systematic processing to encourage information cross-validation, and offered implications for designing trustworthy civic information systems and fostering an accessible, safe, and efficient information-seeking environment in digital space.

cs.HC↗

arXiv2026

Towards Personalized Deep Research: Benchmarks and Evaluations

Deep Research Agents (DRAs) can autonomously conduct complex investigations and generate comprehensive reports, demonstrating strong real-world potential. However, existing evaluations mostly rely on close-ended benchmarks, while open-ended deep research benchmarks remain scarce and typically neglect personalized scenarios. To bridge this gap, we introduce Personalized Deep Research Bench (PDR-Bench), the first benchmark for evaluating personalization in DRAs. It pairs 50 diverse research tasks across 10 domains with 25 authentic user profiles that combine structured persona attributes with dynamic real-world contexts, yielding 250 realistic user-task queries. To assess system performance, we propose the PQR Evaluation Framework, which jointly measures Personalization Alignment, Content Quality, and Factual Reliability. Our experiments on a range of systems highlight current capabilities and limitations in handling personalized deep research. This work establishes a rigorous foundation for developing and evaluating the next generation of truly personalized AI research assistants.

cs.CL↗

arXiv2026

Search More, Think Less: Rethinking Long-Horizon Agentic Search for Efficiency and Generalization

Recent deep research agents primarily improve performance by scaling reasoning depth, but this leads to high inference cost and latency in search-intensive scenarios. Moreover, generalization across heterogeneous research settings remains challenging. In this work, we propose \emph{Search More, Think Less} (SMTL), a framework for long-horizon agentic search that targets both efficiency and generalization. SMTL replaces sequential reasoning with parallel evidence acquisition, enabling efficient context management under constrained context budgets. To support generalization across task types, we further introduce a unified data synthesis pipeline that constructs search tasks spanning both deterministic question answering and open-ended research scenarios with task appropriate evaluation metrics. We train an end-to-end agent using supervised fine-tuning and reinforcement learning, achieving strong and often state of the art performance across benchmarks including BrowseComp (48.6\%), GAIA (75.7\%), Xbench (82.0\%), and DeepResearch Bench (45.9\%). Compared to Mirothinker-v1.0, SMTL with maximum 100 interaction steps reduces the average number of reasoning steps on BrowseComp by 70.7\%, while improving accuracy.

cs.CL↗

arXiv2026

Evaluating Conversational Recommender Systems via Large Language Models: A User-Centric Framework

Conversational recommender systems (CRSs) integrate both recommendation and dialogue tasks, making their evaluation uniquely challenging. Existing approaches primarily assess CRS performance by separately evaluating item recommendation and dialogue management using rule-based metrics. However, these methods fail to capture the real human experience, and they cannot draw direct conclusions about the system's overall performance. As conversational recommender systems become increasingly vital in e-commerce, social media, and customer support, the ability to evaluate both recommendation accuracy and dialogue management quality using a single metric, thereby authentically reflecting user experience, has become the principal challenge impeding progress in this field. In this work, we propose a user-centric evaluation framework based on large language models (LLMs) for CRSs, namely Conversational Recommendation Evaluator (CoRE). CoRE consists of two main components: (1) LLM-As-Evaluator. Firstly, we comprehensively summarize 12 key factors influencing user experience in CRSs and directly leverage LLM as an evaluator to assign a score to each factor. (2) Multi-Agent Debater. Secondly, we design a multi-agent debate framework with four distinct roles (common user, domain expert, linguist, and HCI expert) to discuss and synthesize the 12 evaluation factors into a unified overall performance score. Furthermore, we apply the proposed framework to evaluate four CRSs on two benchmark datasets. The experimental results show that CoRE aligns well with human evaluation in most of the 12 factors and the overall assessment. Especially, CoRE's overall evaluation scores demonstrate significantly better alignment with human feedback compared to existing rule-based metrics.

cs.IR↗

arXiv2026

O-Researcher: An Open Ended Deep Research Model via Multi-Agent Distillation and Agentic RL

The performance gap between closed-source and open-source large language models (LLMs) is largely attributed to disparities in access to high-quality training data. To bridge this gap, we introduce a novel framework for the automated synthesis of sophisticated, research-grade instructional data. Our approach centers on a multi-agent workflow where collaborative AI agents simulate complex tool-integrated reasoning to generate diverse and high-fidelity data end-to-end. Leveraging this synthesized data, we develop a two-stage training strategy that integrates supervised fine-tuning with a novel reinforcement learning method, designed to maximize model alignment and capability. Extensive experiments demonstrate that our framework empowers open-source models across multiple scales, enabling them to achieve new state-of-the-art performance on the major deep research benchmark. This work provides a scalable and effective pathway for advancing open-source LLMs without relying on proprietary data or models.

cs.CL↗

arXiv2026

Judging with Personality and Confidence: A Study on Personality-Conditioned LLM Relevance Assessment

Recent studies have shown that prompting can enable large language models (LLMs) to simulate specific personality traits and produce behaviors that align with those traits. However, there is limited understanding of how these simulated personalities influence critical web search decisions, specifically relevance assessment. Moreover, few studies have examined how simulated personalities impact confidence calibration, specifically the tendencies toward overconfidence or underconfidence. This gap exists even though psychological literature suggests these biases are trait-specific, often linking high extraversion to overconfidence and high neuroticism to underconfidence. To address this gap, we conducted a comprehensive study evaluating multiple LLMs, including commercial models and open-source models, prompted to simulate Big Five personality traits. We tested these models across three test collections (TREC DL 2019, TREC DL 2020, and LLMJudge), collecting two key outputs for each query-document pair: a relevance judgment and a self-reported confidence score. The findings show that personalities such as low agreeableness consistently align more closely with human labels than the unprompted condition. Additionally, low conscientiousness performs well in balancing the suppression of both overconfidence and underconfidence. We also observe that relevance scores and confidence distributions vary systematically across different personalities. Based on the above findings, we incorporate personality-conditioned scores and confidence as features in a random forest classifier. This approach achieves performance that surpasses the best single-personality condition on a new dataset (TREC DL 2021), even with limited training data. These findings highlight that personality-derived confidence offers a complementary predictive signal, paving the way for more reliable and human-aligned LLM evaluators.

cs.CL↗

arXiv2025

O-Mem: Omni Memory System for Personalized, Long Horizon, Self-Evolving Agents

Recent advancements in LLM-powered agents have demonstrated significant potential in generating human-like responses; however, they continue to face challenges in maintaining long-term interactions within complex environments, primarily due to limitations in contextual consistency and dynamic personalization. Existing memory systems often depend on semantic grouping prior to retrieval, which can overlook semantically irrelevant yet critical user information and introduce retrieval noise. In this report, we propose the initial design of O-Mem, a novel memory framework based on active user profiling that dynamically extracts and updates user characteristics and event records from their proactive interactions with agents. O-Mem supports hierarchical retrieval of persona attributes and topic-related context, enabling more adaptive and coherent personalized responses. O-Mem achieves 51.67% on the public LoCoMo benchmark, a nearly 3% improvement upon LangMem,the previous state-of-the-art, and it achieves 62.99% on PERSONAMEM, a 3.5% improvement upon A-Mem,the previous state-of-the-art. O-Mem also boosts token and interaction response time efficiency compared to previous memory frameworks. Our work opens up promising directions for developing efficient and human-like personalized AI assistants in the future.

cs.CL↗

arXiv2025

Chain-of-Agents: End-to-End Agent Foundation Models via Multi-Agent Distillation and Agentic RL

Recent advances in large language models (LLMs) and multi-agent systems have demonstrated remarkable capabilities in complex problem-solving tasks such as deep research, vibe coding, and mathematical reasoning. However, most existing multi-agent systems are built upon manual prompt/workflow engineering with sophisticated agent frameworks, making them computationally inefficient, less capable, and can not benefit from data-centric learning. In this work, we introduce Chain-of-Agents (CoA), a novel paradigm of LLM reasoning that enables native end-to-end complex problem-solving in the same way as a multi-agent system (i.e., multi-turn problem solving with multiple tools and multiple agents) within one model. In chain-of-agents problem-solving, the model dynamically activates different tool agents and role-playing agents to simulate multi-agent collaboration in an end-to-end fashion. To elicit end-to-end chain-of-agents problem-solving abilities in LLMs, we introduce a multi-agent distillation framework to distill state-of-the-art multi-agent systems into chain-of-agents trajectories for agentic supervised fine-tuning. We then use agentic reinforcement learning on verifiable agentic tasks to further improve the models' capabilities on chain-of-agents problem solving. We call the resulting models Agent Foundation Models (AFMs). Our empirical studies demonstrate that AFM establishes new state-of-the-art performance across diverse benchmarks in both web agent and code agent settings. We make the entire research, including the model weights, code for training and evaluation, and the training data, fully open-sourced, which offers a solid starting point for future research on agent models and agentic RL.

cs.AI↗

arXiv2025

"I Always Felt that SomethingWasWrong.": Understanding Compliance Risks and Mitigation Strategies when Highly-Skilled Compliance Knowledge Workers Use Large Language Models

The rapid advancement of Large Language Models (LLMs) has transformed knowledge-intensive has led to its widespread usage by knowledge workers to enhance their productivity. As these professionals handle sensitive information, and the training of text-based GenAI models involves the use of extensive data, there are thus concerns about privacy, security, and broader compliance with regulations and laws. While existing research has addressed privacy and security concerns, the specific compliance risks faced by highly-skilled knowledge workers when using the LLMs, and their mitigation strategies, remain underexplored. As understanding these risks and strategies is crucial for the development of industry-specific compliant LLM mechanisms, this research conducted semi-structured interviews with 24 knowledge workers from knowledge-intensive industries to understand their practices and experiences when integrating LLMs into their workflows. Our research explored how these workers ensure compliance and the resources and challenges they encounter when minimizing risks. Our preliminary findings showed that knowledge workers were concerned about the leakage of sensitive information and took proactive measures such as distorting input data and limiting prompt details to mitigate such risks. Their ability to identify and mitigate risks, however, was significantly hampered by a lack of LLM-specific compliance guidance and training. Our findings highlight the importance of improving knowledge workers' compliance awareness and establishing support systems and compliance cultures within organizations.

cs.HC↗

arXiv2025

SRLAgent: Enhancing Self-Regulated Learning Skills through Gamification and LLM Assistance

Self-regulated learning (SRL) is crucial for college students navigating increased academic demands and independence. Insufficient SRL skills can lead to disorganized study habits, low motivation, and poor time management, undermining learners ability to thrive in challenging environments. Through a formative study involving 59 college students, we identified key challenges students face in developing SRL skills, including difficulties with goal-setting, time management, and reflective learning. To address these challenges, we introduce SRLAgent, an LLM-assisted system that fosters SRL skills through gamification and adaptive support from large language models (LLMs). Grounded in Zimmermans three-phase SRL framework, SRLAgent enables students to engage in goal-setting, strategy execution, and self-reflection within an interactive game-based environment. The system offers real-time feedback and scaffolding powered by LLMs to support students independent study efforts. We evaluated SRLAgent using a between-subjects design, comparing it to a baseline system (SRL without Agent features) and a traditional multimedia learning condition. Results showed significant improvements in SRL skills within the SRLAgent group (p < .001, Cohens d = 0.234) and higher engagement compared to the baselines. This work highlights the value of embedding SRL scaffolding and real-time AI support within gamified environments, offering design implications for educational technologies that aim to promote deeper learning and metacognitive skill development.

cs.HC↗

arXiv2025

DanModCap: Designing a Danmaku Moderation Tool for Video-Sharing Platforms that Leverages Impact Captions with Large Language Models

Online video platforms have gained increased popularity due to their ability to support information consumption and sharing and the diverse social interactions they afford. Danmaku, a real-time commentary feature that overlays user comments on a video, has been found to improve user engagement, however, the use of Danmaku can lead to toxic behaviors and inappropriate comments. To address these issues, we propose a proactive moderation approach inspired by Impact Captions, a visual technique used in East Asian variety shows. Impact Captions combine textual content and visual elements to construct emotional and cognitive resonance. Within the context of this work, Impact Captions were used to guide viewers towards positive Danmaku-related activities and elicit more pro-social behaviors. Leveraging Impact Captions, we developed DanModCap, an moderation tool that collected and analyzed Danmaku and used it as input to large generative language models to produce Impact Captions. Our evaluation of DanModCap demonstrated that Impact Captions reduced negative antagonistic emotions, increased users' desire to share positive content, and elicited self-control in Danmaku social action to fostering proactive community maintenance behaviors. Our approach highlights the benefits of using LLM-supported content moderation methods for proactive moderation in a large-scale live content contexts.

cs.HC↗

arXiv2024

Sportoonizer: Augmenting Sports Highlights' Narration and Visual Impact via Automatic Manga B-Roll Generation

Sports highlights are becoming increasingly popular on video-sharing platforms. Yet, crafting sport highlight videos is challenging, which requires producing engaging narratives from different angles, and conforming to different platform affordances with constantly changing audiences. Many content creators therefore create derivative work of the original sports video through manga styles to enhance its expressiveness. But manually creating and inserting tailored manga-style content can still be time-consuming. We introduce Sportoonizer, a system embedding the pipeline for automatic generation of manga-style animations for highlights in sports videos and insertion into original videos. It seamlessly merges dynamic manga sequences with live-action footage, enriching the visual tapestry and deepening narrative scope. By leveraging genAIs, Sportoonizer crafts compelling storylines encapsulating the intensity of sports moments and athletes' personal journeys. Our evaluation study demonstrates that integrating manga B-rolls significantly enhances viewer engagement, visual interest, and emotional connection towards athletes' stories in the viewing experience.

cs.HC↗

arXiv2024

"It Felt Like Having a Second Mind": Investigating Human-AI Co-creativity in Prewriting with Large Language Models

Prewriting is the process of discovering and developing ideas before a first draft, which requires divergent thinking and often implies unstructured strategies such as diagramming, outlining, free-writing, etc. Although large language models (LLMs) have been demonstrated to be useful for a variety of tasks including creative writing, little is known about how users would collaborate with LLMs to support prewriting. The preferred collaborative role and initiative of LLMs during such a creativity process is also unclear. To investigate human-LLM collaboration patterns and dynamics during prewriting, we conducted a three-session qualitative study with 15 participants in two creative tasks: story writing and slogan writing. The findings indicated that during collaborative prewriting, there appears to be a three-stage iterative Human-AI Co-creativity process that includes Ideation, Illumination, and Implementation stages. This collaborative process champions the human in a dominant role, in addition to mixed and shifting levels of initiative that exist between humans and LLMs. This research also reports on collaboration breakdowns that occur during this process, user perceptions of using existing LLMs during Human-AI Co-creativity, and discusses design implications to support this co-creativity process.

cs.HC↗

arXiv2023

Let's Play Together through Channels: Understanding the Practices and Experience of Danmaku Participation Game Players in China

Live streaming is becoming increasingly popular in recent years, as most channels prioritize the delivery of engaging content to their viewers. Among various live streaming channels, Danmaku participation game (DPG) has emerged in China as a mixture of live streaming and online gaming, offering an immersive gaming experience to players. Although prior research has explored audience participation games (APGs) in North America and Europe, it primarily focuses on discussing prototypes and lacks observation of players in natural settings. Little is known about how players perceive DPGs and their player experience. To fill the research gap, we observed a series of DPG channels and conducted an interview-based study to gain insights into the practices and experiences of DPG players. Our work reveals that DPGs can effectively synergize live streaming and online games, amplifying both player engagement and a profound sense of accomplishment to players.

cs.HC↗