Weiwen Liu

Jianghao Lin, Xinyi Dai, Yunjia Xi et al.

With the rapid development of online services, recommender systems (RS) have become increasingly indispensable for mitigating information overload. Despite remarkable progress, conventional recommendation models (CRM) still have some limitations, e.g., lacking open-world knowledge, and difficulties in comprehending users' underlying preferences and motivations. Meanwhile, large language models (LLM) have shown impressive general intelligence and human-like capabilities, which mainly stem from their extensive open-world knowledge, reasoning ability, as well as their comprehension of human culture and society. Consequently, the emergence of LLM is inspiring the design of recommender systems and pointing out a promising research direction, i.e., whether we can incorporate LLM and benefit from their knowledge and capabilities to compensate for the limitations of CRM. In this paper, we conduct a comprehensive survey on this research direction from the perspective of the whole pipeline in real-world recommender systems. Specifically, we summarize existing works from two orthogonal aspects: where and how to adapt LLM to RS. For the WHERE question, we discuss the roles that LLM could play in different stages of the recommendation pipeline, i.e., feature engineering, feature encoder, scoring/ranking function, user interaction, and pipeline controller. For the HOW question, we investigate the training and inference strategies, resulting in two fine-grained taxonomy criteria, i.e., whether to tune LLM or not, and whether to involve conventional recommendation models for inference. Then, we highlight key challenges in adapting LLM to RS from three aspects, i.e., efficiency, effectiveness, and ethics. Finally, we summarize the survey and discuss the future prospects. We actively maintain a GitHub repository for papers and other related resources: https://github.com/CHIANGEL/Awesome-LLM-for-RecSys/.

Weiwen Liu, Xu Huang, Xingshan Zeng et al.

Function calling significantly extends the application boundary of large language models, where high-quality and diverse training data is critical for unlocking this capability. However, real function-calling data is quite challenging to collect and annotate, while synthetic data generated by existing pipelines tends to lack coverage and accuracy. In this paper, we present ToolACE, an automatic agentic pipeline designed to generate accurate, complex, and diverse tool-learning data. ToolACE leverages a novel self-evolution synthesis process to curate a comprehensive API pool of 26,507 diverse APIs. Dialogs are further generated through the interplay among multiple agents, guided by a formalized thinking process. To ensure data accuracy, we implement a dual-layer verification system combining rule-based and model-based checks. We demonstrate that models trained on our synthesized data, even with only 8B parameters, achieve state-of-the-art performance on the Berkeley Function-Calling Leaderboard, rivaling the latest GPT-4 models. Our model and a subset of the data are publicly available at https://huggingface.co/Team-ACE.

Wentao Ning, Xiao Yan, Weiwen Liu et al.

Multi-domain recommendation (MDR) aims to provide recommendations for different domains (e.g., types of products) with overlapping users/items and is common for platforms such as Amazon, Facebook, and LinkedIn that host multiple services. Existing MDR models face two challenges: First, it is difficult to disentangle knowledge that generalizes across domains (e.g., a user likes cheap items) and knowledge specific to a single domain (e.g., a user likes blue clothing but not blue cars). Second, they have limited ability to transfer knowledge across domains with small overlaps. We propose a new MDR method named EDDA with two key components, i.e., embedding disentangling recommender and domain alignment, to tackle the two challenges respectively. In particular, the embedding disentangling recommender separates both the model and embedding for the inter-domain part and the intra-domain part, while most existing MDR methods only focus on model-level disentangling. The domain alignment leverages random walks from graph processing to identify similar user/item pairs from different domains and encourages similar user/item pairs to have similar embeddings, enhancing knowledge transfer. We compare EDDA with 12 state-of-the-art baselines on 3 real datasets. The results show that EDDA consistently outperforms the baselines on all datasets and domains. All datasets and codes are available at https://github.com/Stevenn9981/EDDA.

Jianghao Lin, Weiwen Liu, Xinyi Dai et al.

To better exploit search logs and model users' behavior patterns, numerous click models are proposed to extract users' implicit interaction feedback. Most traditional click models are based on the probabilistic graphical model (PGM) framework, which requires manually designed dependencies and may oversimplify user behaviors. Recently, methods based on neural networks are proposed to improve the prediction accuracy of user behaviors by enhancing the expressive ability and allowing flexible dependencies. However, they still suffer from the data sparsity and cold-start problems. In this paper, we propose a novel graph-enhanced click model (GraphCM) for web search. Firstly, we regard each query or document as a vertex, and propose novel homogeneous graph construction methods for queries and documents respectively, to fully exploit both intra-session and inter-session information for the sparsity and cold-start problems. Secondly, following the examination hypothesis, we separately model the attractiveness estimator and examination predictor to output the attractiveness scores and examination probabilities, where graph neural networks and neighbor interaction techniques are applied to extract the auxiliary information encoded in the pre-constructed homogeneous graphs. Finally, we apply combination functions to integrate examination probabilities and attractiveness scores into click predictions. Extensive experiments conducted on three real-world session datasets show that GraphCM not only outperforms the state-of-art models, but also achieves superior performance in addressing the data sparsity and cold-start problems.

Lingyue Fu, Jianghao Lin, Weiwen Liu et al.

To provide click simulation or relevance estimation based on users' implicit interaction feedback, click models have been much studied during recent years. Most click models focus on user behaviors towards a single list. However, with the development of user interface (UI) design, the layout of displayed items on a result page tends to be multi-block (i.e., multi-list) style instead of a single list, which requires different assumptions to model user behaviors more accurately. There exist click models for multi-block pages in desktop contexts, but they cannot be directly applied to mobile scenarios due to different interaction manners, result types and especially multi-block presentation styles. In particular, multi-block mobile pages can normally be decomposed into interleavings of basic vertical blocks and horizontal blocks, thus resulting in typically F-shape forms. To mitigate gaps between desktop and mobile contexts for multi-block pages, we conduct a user eye-tracking study, and identify users' sequential browsing, block skip and comparison patterns on F-shape pages. These findings lead to the design of a novel F-shape Click Model (FSCM), which serves as a general solution to multi-block mobile pages. Firstly, we construct a directed acyclic graph (DAG) for each page, where each item is regarded as a vertex and each edge indicates the user's possible examination flow. Secondly, we propose DAG-structured GRUs and a comparison module to model users' sequential (sequential browsing, block skip) and non-sequential (comparison) behaviors respectively. Finally, we combine GRU states and comparison patterns to perform user click predictions. Experiments on a large-scale real-world dataset validate the effectiveness of FSCM on user behavior predictions compared with baseline models.

Zhicheng He, Weiwen Liu, Wei Guo et al.

User Behavior Modeling (UBM) plays a critical role in user interest learning, which has been extensively used in recommender systems. Crucial interactive patterns between users and items have been exploited, which brings compelling improvements in many recommendation tasks. In this paper, we attempt to provide a thorough survey of this research topic. We start by reviewing the research background of UBM. Then, we provide a systematic taxonomy of existing UBM research works, which can be categorized into four different directions including Conventional UBM, Long-Sequence UBM, Multi-Type UBM, and UBM with Side Information. Within each direction, representative models and their strengths and weaknesses are comprehensively discussed. Besides, we elaborate on the industrial practices of UBM methods with the hope of providing insights into the application value of existing UBM solutions. Finally, we summarize the survey and discuss the future prospects of this field.

Zicai Cui, Zihan Guo, Weiwen Liu et al.

Skill-based LLM agents increasingly rely on long procedural documents, but full-document prompting wastes tokens and dilutes information critical to execution. We study this setting as intra-skill retrieval, where the goal is to select a minimal, execution-sufficient context from a known skill document given a query. We present SkillPager, a two-stage framework that parses each Markdown skill into typed semantic nodes offline and leverages Maximal Marginal Relevance (MMR) to perform global, query-conditioned node selection online. On a benchmark of 395 skills and 1,975 queries, SkillPager achieves 78.89% LLM-judged context sufficiency, compared to 82.23% for the exhaustive full-document baseline, while reducing prompt tokens by 47.04%. A granularity ablation shows that applying the same retrieval algorithm to raw fixed-length chunks reaches a comparable 81.77% sufficiency but increases token cost by 28.81%, demonstrating that efficiency gains are driven by typed semantic granularity rather than the retrieval algorithm alone. Among graph-based baselines, SkillPager outperforms the strongest baseline by a margin of 12.16%. Further ablations show that supporting content is most effective when retained in the candidate pool and selected adaptively rather than removed by static heuristics. These results identify typed intra-document retrieval as a distinct access problem for skill-based agents.

Yi Li, Jieming Zhu, Weiwen Liu et al.

The goal of recommender systems is to provide ordered item lists to users that best match their interests. As a critical task in the recommendation pipeline, re-ranking has received increasing attention in recent years. In contrast to conventional ranking models that score each item individually, re-ranking aims to explicitly model the mutual influences among items to further refine the ordering of items given an initial ranking list. In this paper, we present a personalized re-ranking model (dubbed PEAR) based on contextualized transformer. PEAR makes several major improvements over the existing methods. Specifically, PEAR not only captures feature-level and item-level interactions, but also models item contexts from both the initial ranking list and the historical clicked item list. In addition to item-level ranking score prediction, we also augment the training of PEAR with a list-level classification task to assess users' satisfaction on the whole ranking list. Experimental results on both public and production datasets have shown the superior effectiveness of PEAR compared to the previous re-ranking models.

Siyuan Qi, Xinyuan Wang, Yingxuan Yang et al.

Agentic Retrieval-Augmented Generation improves retrieval by integrating planning, tool use, and iterative reasoning, but existing agentic RAG methods often couple semantic expansion with retrieval decisions in short-horizon inference loops, leading to high inference cost and limited suitability for time-sensitive news retrieval. We propose DynaTree, a two-stage framework for efficient and adaptive news retrieval. In the offline stage, DynaTree uses coordinated agents to construct a reusable retrieval tree that materializes the semantic space of a query topic. In the online stage, DynaTree performs lightweight daily subtree selection over a time-localized evaluation proxy, without further agentic reasoning, tree modification, or retraining. Experiments on a multi-day Syft news benchmark and multiple BEIR datasets show that DynaTree achieves strong recall and ranking performance, consistently outperforming standard RAG and prior agentic baselines. We further deploy DynaTree in the Syft production system and evaluate it through online A/B testing from Jan. 28 to Feb. 6, 2026. The dynamically adapted variant improves survival rate from 0.32-0.53 to 0.59-0.73 over a fixed offline-selected subtree and outperforms existing production recallers on every evaluation day. These results show that persistent, structure-aware semantic expansion can translate offline agentic reasoning into practical improvements in coverage, freshness, and relevance for real-world news retrieval.

Kangning Zhang, Wenxiang Jiao, Kounianhua Du et al.

Augmenting Large Language Models (LLMs) with external tools enables them to execute complex, multi-step tasks. However, tool learning is hampered by the static synthetic data pipelines where data generation and model training are executed as two separate, non-interactive processes. This approach fails to adaptively focus on a model's specific weaknesses and allows noisy labels to persist, degrading training efficiency. We introduce LoopTool, a fully automated, model-aware data evolution framework that closes this loop by tightly integrating data synthesis and model training. LoopTool iteratively refines both the data and the model through three synergistic modules: (1) Greedy Capability Probing (GCP) diagnoses the model's mastered and failed capabilities; (2) Judgement-Guided Label Verification (JGLV) uses an open-source judge model to find and correct annotation errors, progressively purifying the dataset; and (3) Error-Driven Data Expansion (EDDE) generates new, challenging samples based on identified failures. This closed-loop process operates within a cost-effective, open-source ecosystem, eliminating dependence on expensive closed-source APIs. Experiments show that our 8B model trained with LoopTool significantly surpasses its 32B data generator and achieves new state-of-the-art results on the BFCL-v3 and ACEBench benchmarks for its scale. Our work demonstrates that closed-loop, self-refining data pipelines can dramatically enhance the tool-use capabilities of LLMs.

Kounianhua Du, Jianxing Liu, Kangning Zhang et al.

The rapid evolution of large language models (LLMs) has intensified the demand for effective personalization techniques that can adapt model behavior to individual user preferences. Despite the non-parametric methods utilizing the in-context learning ability of LLMs, recent parametric adaptation methods, including personalized parameter-efficient fine-tuning and reward modeling emerge. However, these methods face limitations in handling dynamic user patterns and high data sparsity scenarios, due to low adaptability and data efficiency. To address these challenges, we propose a fine-grained and instance-tailored steering framework that dynamically generates sample-level interference vectors from user data and injects them into the model's forward pass for personalized adaptation. Our approach introduces two key technical innovations: a fine-grained steering component that captures nuanced signals by hooking activations from attention and MLP layers, and an input-aware aggregation module that synthesizes these signals into contextually relevant enhancements. The method demonstrates high flexibility and data efficiency, excelling in fast-changing distribution and high data sparsity scenarios. In addition, the proposed method is orthogonal to existing methods and operates as a plug-in component compatible with different personalization techniques. Extensive experiments across diverse scenarios--including short-to-long text generation, and web function calling--validate the effectiveness and compatibility of our approach. Results show that our method significantly enhances personalization performance in fast-shifting environments while maintaining robustness across varying interaction modes and context lengths. Implementation is available at https://github.com/KounianhuaDu/Fints.

Huacan Chai, Yukai Wang, Yingxuan Yang et al.

Existing benchmarks for multimodal memory reasoning largely evaluate systems within pre-assembled contexts, but under-evaluate whether agents can use evidence distributed across independently originated sources. We argue that source-distributed memory composition is an important and under-examined bottleneck in multimodal agent memory, especially when relevant evidence is fragmented across heterogeneous artifacts such as conversations, profiles, screenshots, tables, images, and documents. To address this gap, we introduce Source-distributed Multimodal Memory Benchmark(SMMBench), which measures whether agents can retrieve, align, and compose multimodal evidence scattered across multiple sources rather than reason within a single curated context. SMMBench evaluates four core capabilities: (1) cross-source multimodal reasoning; (2) conflict resolution; (3) preference reasoning; (4) memory-grounded action prediction. The benchmark contains 1877 samples grounded in 264 sources. Experiments on representative memory-style and retrieval-based baselines show that current systems still struggle on these capabilities, positioning source-distributed multimodal memory as an important and still under-evaluated challenge for multimodal agents. Our data are available at https://huggingface.co/datasets/HuacanChai/SMMBench.

Zheng Wu, Xingyu Lou, Xinbei Ma et al.

Large Language Model (LLM)-based agents significantly extend the utility of LLMs by interacting with dynamic environments. However, enabling agents to continually learn new tasks without catastrophic forgetting remains a critical challenge, known as the stability-plasticity dilemma. In this work, we argue that this dilemma fundamentally arises from the failure to explicitly distinguish between common knowledge shared across tasks and conflicting knowledge introduced by task-specific interference. To address this, we propose Agent-Dice, a parameter fusion framework based on directional consensus evaluation. Concretely, Agent-Dice disentangles knowledge updates through a two-stage process: geometric consensus filtering to prune conflicting gradients, and curvature-based importance weighting to amplify shared semantics. We provide a rigorous theoretical analysis that establishes the validity of the proposed fusion scheme and offers insight into the origins of the stability-plasticity dilemma. Extensive experiments on GUI agents and tool-use agent domains demonstrate that Agent-Dice exhibits outstanding continual learning performance with minimal computational overhead and parameter updates. The codes are available at https://github.com/Wuzheng02/Agent-Dice.

Haoyi Hu, Qirong Lyu, Xianghan Kong et al.

While AI agents demonstrate remarkable capabilities in reasoning and tool use, they remain fundamentally reactive: they compute responses only after explicit user prompts. This paradigm ignores a critical opportunity: the idle time between interactions is largely wasted, leaving agents unable to prepare for future user needs. To bridge this gap, we introduce ProAct, a proactive agent architecture that leverages idle-time compute to anticipate and fulfill likely upcoming user needs. By analyzing evolving dialogue history together with persistent memory, ProAct predicts upcoming needs and iteratively acquires information, allowing the agent to resolve knowledge gaps and prepare evidence before the user initiates a query.To rigorously evaluate proactive capabilities, we also introduce ProActEval, a comprehensive benchmark comprising 200 scenarios across 40 domains, featuring predictable need chains and diverse user cognitive profiles. Empirical results demonstrate significant advantages over reactive baselines. ProAct accelerates task completion by reducing required turns by 14.8%, decreases user effort by 11.7%, and cuts hallucination rates by 28.1% on ProActEval. Furthermore, MemBench evaluations confirm that ProAct achieves state-of-the-art reflective accuracy, underscoring its sustained and robust performance.

Xiaohang Nie, Zihan Guo, Kezhuo Yang et al.

AI agents are rapidly expanding in both capability and population: they now write code, operate computers across platforms, manage cloud infrastructure, and make purchasing decisions, while open-source frameworks such as OpenClaw are putting personal agents in the hands of millions and embodied agents are spreading across smartphones, vehicles, and robots. As the internet prepares to host billions of such entities, it is shifting toward what we call Open Agentic Web, a decentralized digital ecosystem in which agents from different users, organizations, and runtimes can discover one another, negotiate task boundaries, and delegate work across open technical and social surfaces at scale. Yet most of today's agents remain isolated tools or closed-ecosystem orchestrators rather than socially integrated participants in open networks. We argue that the next generation of agents must become Agentic Citizens, defined by three requirements: Agentic-Web-Native Collaboration, participation in open collaboration networks rather than only closed internal orchestration; Agent Identity and Personhood, continuity as a social entity rather than a resettable function call; and Lifelong Evolution, improvement across task performance, communication, and collaboration over time. We present Synergy, a general-purpose agent architecture and runtime harness for persistent, collaborative, and evolving agents on Open Agentic Web, grounding collaboration in session-native orchestration, repository-backed workspaces, and social communication; identity in typed memory, notes, agenda, skills, and persistent social relationships; and evolution in an experience-centered learning mechanism that proactively recalls rewarded trajectories at inference time.

Zhenhua Dong, Jieming Zhu, Weiwen Liu et al.

Huawei's vision and mission is to build a fully connected intelligent world. Since 2013, Huawei Noah's Ark Lab has helped many products build recommender systems and search engines for getting the right information to the right users. Every day, our recommender systems serve hundreds of millions of mobile phone users and recommend different kinds of content and services such as apps, news feeds, songs, videos, books, themes, and instant services. The big data and various scenarios provide us with great opportunities to develop advanced recommendation technologies. Furthermore, we have witnessed the technical trend of recommendation models in the past ten years, from the shallow and simple models like collaborative filtering, linear models, low rank models to deep and complex models like neural networks, pre-trained language models. Based on the mission, opportunities and technological trends, we have also met several hard problems in our recommender systems. In this talk, we will share ten important and interesting challenges and hope that the RecSys community can get inspired and create better recommender systems.

Shijue Huang, Wanjun Zhong, Jianqiao Lu et al.

The recent trend of using Large Language Models (LLMs) as tool agents in real-world applications underscores the necessity for comprehensive evaluations of their capabilities, particularly in complex scenarios involving planning, creating, and using tools. However, existing benchmarks typically focus on simple synthesized queries that do not reflect real-world complexity, thereby offering limited perspectives in evaluating tool utilization. To address this issue, we present UltraTool, a novel benchmark designed to improve and evaluate LLMs' ability in tool utilization within real-world scenarios. UltraTool focuses on the entire process of using tools - from planning and creating to applying them in complex tasks. It emphasizes real-world complexities, demanding accurate, multi-step planning for effective problem-solving. A key feature of UltraTool is its independent evaluation of planning with natural language, which happens before tool usage and simplifies the task solving by mapping out the intermediate steps. Thus, unlike previous work, it eliminates the restriction of pre-defined toolset. Through extensive experiments on various LLMs, we offer novel insights into the evaluation of capabilities of LLMs in tool utilization, thereby contributing a fresh perspective to this rapidly evolving field. The benchmark is publicly available at https://github.com/JoeYing1019/UltraTool.

Yaxi Lu, Shenzhi Yang, Cheng Qian et al. · tsinghua

Agents powered by large language models have shown remarkable abilities in solving complex tasks. However, most agent systems remain reactive, limiting their effectiveness in scenarios requiring foresight and autonomous decision-making. In this paper, we tackle the challenge of developing proactive agents capable of anticipating and initiating tasks without explicit human instructions. We propose a novel data-driven approach for this problem. Firstly, we collect real-world human activities to generate proactive task predictions. These predictions are then labeled by human annotators as either accepted or rejected. The labeled data is used to train a reward model that simulates human judgment and serves as an automatic evaluator of the proactiveness of LLM agents. Building on this, we develop a comprehensive data generation pipeline to create a diverse dataset, ProactiveBench, containing 6,790 events. Finally, we demonstrate that fine-tuning models with the proposed ProactiveBench can significantly elicit the proactiveness of LLM agents. Experimental results show that our fine-tuned model achieves an F1-Score of 66.47% in proactively offering assistance, outperforming all open-source and close-source models. These results highlight the potential of our method in creating more proactive and effective agent systems, paving the way for future advancements in human-agent collaboration.

Jihong Wang, Jiamu Zhou, Weiming Zhang et al.

With the advancement of vision-language models, web automation has made significant progress. However, deploying autonomous agents in real-world settings remains challenging, primarily due to site heterogeneity, where generalist models lack domain-specific priors for diverse interfaces, and long-horizon instability, characterized by the accumulation of decision drift over extended interactions. To address these challenges, we introduce ColorBrowserAgent (Complex Long-Horizon Browser Agent), a knowledge-evolving agent for robust web automation. Our approach addresses these challenges through two synergistic mechanisms: human-in-the-loop knowledge adaptation that transforms sparse human feedback into reusable domain knowledge, and knowledge-aligned progressive summarization that stabilizes long interactions through memory compression. Extensive experiments on WebArena, WebChoreArena and industrial deployment show that ColorBrowserAgent consistently outperforms strong baselines. It achieves a state-of-the-art success rate of 71.2% on WebArena and maintains 47.4% performance under zero-shot transfer setting on WebChoreArena. In commercial deployment, it improves user satisfaction by 19.3% relatively, verifying its robustness in real-world scenarios.

Xingshan Zeng, Lingzhi Wang, Weiwen Liu et al.

Current test-time scaling (TTS) techniques enhance large language model (LLM) performance by allocating additional computation at inference time, yet they remain insufficient for agentic settings, where actions directly interact with external environments and their effects can be irreversible and costly. We propose ARTIS, Agentic Risk-Aware Test-Time Scaling via Iterative Simulation, a framework that decouples exploration from commitment by enabling test-time exploration through simulated interactions prior to real-world execution. This design allows extending inference-time computation to improve action-level reliability and robustness without incurring environmental risk. We further show that naive LLM-based simulators struggle to capture rare but high-impact failure modes, substantially limiting their effectiveness for agentic decision making. To address this limitation, we introduce a risk-aware tool simulator that emphasizes fidelity on failure-inducing actions via targeted data generation and rebalanced training. Experiments on multi-turn and multi-step agentic benchmarks demonstrate that iterative simulation substantially improves agent reliability, and that risk-aware simulation is essential for consistently realizing these gains across models and tasks.

Jiachen Zhu, Lingyu Yang, Rong Shan et al.

The rise of autonomous GUI agents has triggered adversarial countermeasures from digital platforms, yet existing research prioritizes utility and robustness over the critical dimension of anti-detection. We argue that for agents to survive in human-centric ecosystems, they must evolve Humanization capabilities. We introduce the ``Turing Test on Screen,'' formally modeling the interaction as a MinMax optimization problem between a detector and an agent aiming to minimize behavioral divergence. We then collect a new high-fidelity dataset of mobile touch dynamics, and conduct our analysis that vanilla LMM-based agents are easily detectable due to unnatural kinematics. Consequently, we establish the Agent Humanization Benchmark (AHB) and detection metrics to quantify the trade-off between imitability and utility. Finally, we propose methods ranging from heuristic noise to data-driven behavioral matching, demonstrating that agents can achieve high imitability theoretically and empirically without sacrificing performance. This work shifts the paradigm from whether an agent can perform a task to how it performs it within a human-centric ecosystem, laying the groundwork for seamless coexistence in adversarial digital environments.

Junjie Huang, Jiarui Qin, Di Yin et al.

Standard training pipelines for large language models (LLMs) are typically unidirectional, progressing from pre-training to post-training. However, the potential for a bidirectional process--where insights from post-training retroactively improve the pre-trained foundation--remains unexplored. We aim to establish a self-reinforcing flywheel: a cycle in which reinforcement learning (RL)-tuned model strengthens the base model, which in turn enhances subsequent post-training performance, requiring no specially trained teacher or reference model. To realize this, we analyze training dynamics and identify the mid-training (annealing) phase as a critical turning point for model capabilities. This phase typically occurs at the end of pre-training, utilizing high-quality corpora under a rapidly decaying learning rate. Building upon this insight, we introduce ReMiT (Reinforcement Learning-Guided Mid-Training). Specifically, ReMiT leverages the reasoning priors of RL-tuned models to dynamically reweight tokens during the mid-training phase, prioritizing those pivotal for reasoning. Empirically, ReMiT achieves an average improvement of 3\% on 10 pre-training benchmarks, spanning math, code, and general reasoning, and sustains these gains by over 2\% throughout the post-training pipeline. These results validate an iterative feedback loop, enabling continuous and self-reinforcing evolution of LLMs.

Yunjia Xi, Jianghao Lin, Yongzhao Xiao et al.

The advent of Large Language Models (LLMs) has significantly revolutionized web search. The emergence of LLM-based Search Agents marks a pivotal shift towards deeper, dynamic, autonomous information seeking. These agents can comprehend user intentions and environmental context and execute multi-turn retrieval with dynamic planning, extending search capabilities far beyond the web. Leading examples like OpenAI's Deep Research highlight their potential for deep information mining and real-world applications. This survey provides the first systematic analysis of search agents. We comprehensively analyze and categorize existing works from the perspectives of architecture, optimization, application, and evaluation, ultimately identifying critical open challenges and outlining promising future research directions in this rapidly evolving field. Our repository is available on https://github.com/YunjiaXi/Awesome-Search-Agent-Papers.

Xiaohang Nie, Zihan Guo, Zicai Cui et al.

As large language models (LLM)-driven agents transition from isolated task solvers to persistent digital entities, the emergence of the Agentic Web, an ecosystem where heterogeneous agents autonomously interact and co-evolve, marks a pivotal shift toward Artificial General Intelligence (AGI). However, LLM-based multi-agent systems (LaMAS) are hindered by open-world issues such as scaling friction, coordination breakdown, and value dissipation. To address these challenges, we introduce Holos, a web-scale LaMAS architected for long-term ecological persistence. Holos adopts a five-layer architecture, with core modules primarily featuring the Nuwa engine for high-efficiency agent generation and hosting, a market-driven Orchestrator for resilient coordination, and an endogenous value cycle to achieve incentive compatibility. By bridging the gap between micro-level collaboration and macro-scale emergence, Holos hopes to lay the foundation for the next generation of the self-organizing and continuously evolving Agentic Web. We have publicly released Holos (accessible at https://holosai.io), providing a resource for the community and a testbed for future research in large-scale agentic ecosystems.

Lingyue Fu, Xin Ding, Yaoming Zhu et al.

Large Language Model (LLM) agents have evolved from basic text generation to autonomously completing complex tasks through interaction with external tools. However, current benchmarks mainly assess end-to-end performance in fixed scenarios, restricting evaluation to specific skills and suffering from score saturation and growing dependence on expert annotation as agent capabilities improve. In this work, we emphasize the importance of learning ability, including both self-improvement and peer-learning, as a core driver for agent evolution toward human-level intelligence. We propose an iterative, competitive peer-learning framework, which allows agents to refine and optimize their strategies through repeated interactions and feedback, thereby systematically evaluating their learning capabilities. To address the score saturation issue in current benchmarks, we introduce CATArena, a tournament-style evaluation platform featuring four diverse board and card games with open-ended scoring. By providing tasks without explicit upper score limits, CATArena enables continuous and dynamic evaluation of rapidly advancing agent capabilities. Experimental results and analyses involving both minimal and commercial code agents demonstrate that CATArena provides reliable, stable, and scalable benchmarking for core agent abilities, particularly learning ability and strategy coding.

Yirong Zeng, Xiao Ding, Yutai Hou et al.

Training tool-augmented LLMs has emerged as a promising approach to enhancing language models' capabilities for complex tasks. The current supervised fine-tuning paradigm relies on constructing extensive domain-specific datasets to train models. However, this approach often struggles to generalize effectively to unfamiliar or intricate tool-use scenarios. Recently, reinforcement learning (RL) paradigm can endow LLMs with superior reasoning and generalization abilities. In this work, we address a key question: Can the pure RL be used to effectively elicit a model's intrinsic reasoning capabilities and enhance the tool-agnostic generalization? We propose a dynamic generalization-guided reward design for rule-based RL, which progressively shifts rewards from exploratory to exploitative tool-use patterns. Based on this design, we introduce the Tool-Zero series models. These models are trained to enable LLMs to autonomously utilize general tools by directly scaling up RL from Zero models (i.e., base models without post-training). Experimental results demonstrate that our models achieve over 7% performance improvement compared to both SFT and RL-with-SFT models under the same experimental settings. These gains are consistently replicated across cross-dataset and intra-dataset evaluations, validating the effectiveness and robustness of our methods.

Congmin Zheng, Xiaoyun Mo, Xinbei Ma et al.

Reinforcement Learning (RL) has emerged as a mainstream paradigm for training Mobile GUI Agents, yet it struggles with the temporal credit assignment problem inherent in long-horizon tasks. A primary challenge lies in the trade-off between reward fidelity and density: outcome reward offers high fidelity but suffers from signal sparsity, while process reward provides dense supervision but remains prone to bias and reward hacking. To resolve this conflict, we propose the Adaptive Milestone Reward (ADMIRE) mechanism. ADMIRE constructs a verifiable, adaptive reward system by anchoring trajectory to milestones, which are dynamically distilled from successful explorations. Crucially, ADMIRE integrates an asymmetric credit assignment strategy that denoises successful trajectories and scaffolds failed trajectories. Extensive experiments demonstrate that ADMIRE consistently yields over 10% absolute improvement in success rate across different base models on AndroidWorld. Moreover, the method exhibits robust generalizability, achieving strong performance across diverse RL algorithms and heterogeneous environments such as web navigation and embodied tasks.

Jingxing Wang, Chenyu Zhou, Zhihui Fu et al.

LLM agents benefit from reusable skills, yet test-time tasks often require guidance more specific than a static skill library can provide. We propose \emph{SkillTTA}, a Test-Time Adaptive Skill Synthesis method that retrieves a small set of training trajectories relevant to the current task and synthesizes them into a temporary, task-specific textual skill. The solver model is kept fixed, so adaptation happens entirely through generated context rather than parameter updates. We evaluate the method on SpreadsheetBench, ALFWorld, and BigCodeBench. Compared with static trajectory-to-skill synthesis using GPT-5.5, task-specific skills improve SpreadsheetBench Pass@1 from 0.397 to 0.505 and BigCodeBench Pass@1 from 0.517 to 0.651. On ALFWorld, the method matches a heavier memory-learning baseline within four points of success rate while producing the shortest successful trajectories among reported methods. Ablations on SpreadsheetBench further show that synthesized skills outperform raw trajectory prompting, that top-$k$ retrieval should stay small, and that failed trajectories are especially useful because they expose recurring evaluator-facing mistakes.

Kangning Zhang, Shuai Shao, Qingyao Li et al.

Reusable skills have become a core substrate for improving agent capabilities, yet most existing skill packages encode reusable behavior primarily as textual prompts, executable code, or learned routines. For visual agents, however, procedural knowledge is inherently multimodal: reuse depends not only on what operation to perform, but also on recognizing the relevant state, interpreting visual evidence of progress or failure, and deciding what to do next. We formalize this requirement as multimodal procedural knowledge and address three practical challenges: (I) what a multimodal skill package should contain; (II) where such packages can be derived from public interaction experience; and (III) how agents can consult multimodal evidence at inference time without excessive image context or over-anchoring to reference screenshots. We introduce MMSkills, a framework for representing, generating, and using reusable multimodal procedures for runtime visual decision making. Each MMSkill is a compact, state-conditioned package that couples a textual procedure with runtime state cards and multi-view keyframes. To construct these packages, we develop an agentic trajectory-to-skill Generator that transforms public non-evaluation trajectories into reusable multimodal skills through workflow grouping, procedure induction, visual grounding, and meta-skill-guided auditing. To use them, we introduce a branch-loaded multimodal skill agent: selected state cards and keyframes are inspected in a temporary branch, aligned with the live environment, and distilled into structured guidance for the main agent. Experiments across GUI and game-based visual-agent benchmarks show that MMSkills consistently improve both frontier and smaller multimodal agents, suggesting that external multimodal procedural knowledge complements model-internal priors.

Zheng Wu, Heyuan Huang, Xingyu Lou et al.

With the rapid progress of multimodal large language models, operating system (OS) agents become increasingly capable of automating tasks through on-device graphical user interfaces (GUIs). However, most existing OS agents are designed for idealized settings, whereas real-world environments often present untrustworthy conditions. To mitigate risks of over-execution in such scenarios, we propose a query-driven human-agent-GUI interaction framework that enables OS agents to decide when to query humans for more reliable task completion. Built upon this framework, we introduce VeriOS-Agent, a trustworthy OS agent trained with a two-stage learning paradigm that falicitate the decoupling and utilization of meta-knowledge. Concretely, VeriOS-Agent autonomously executes actions in normal conditions while proactively querying humans in untrustworthy scenarios. Experiments show that VeriOS-Agent improves the average step-wise success rate by 20.64\% in untrustworthy scenarios over the state-of-the-art, without compromising normal performance. Analysis highlights VeriOS-Agent's rationality, generalizability, and scalability. The codes, datasets and models are available at https://github.com/Wuzheng02/VeriOS.

Yirong Zeng, Xiao Ding, Yuxian Wang et al.

Augmenting large language models (LLMs) with external tools is a promising approach to enhance their capabilities, especially for complex tasks. Synthesizing tool-use data through real-world simulations is an effective way to achieve this. However, our investigation reveals that training gains significantly decay as synthetic data increases. The model struggles to benefit from additional synthetic data, which fails to endow it with advanced tool-use capabilities in complex scenarios Moreover, we discovered that the above limitation usually manifests as a fragment deficiency (i.e., parameter errors) in response. To this end, we propose an iterative reinforced fine-tuning strategy designed to alleviate this limitation. This strategy involves: (1) enhancing the diversity of response for synthetic data through path exploration of Monte Carlo Tree Search. (2) iteratively pinpointing the model's deficiency by constructing fine-grained preference pairs, and then improving it by preference optimization algorithms for targeted improvement. The experiments show that our method achieves 13.11% better performance than the same-size base model. It achieves an improvement of 6.5% in complex scenarios compared to the baseline, and it also outperforms larger open-source and closed-source models.

Jiachen Zhu, Zhuoying Ou, Congmin Zheng et al.

Large Language Models (LLMs) are highly sensitive to their input contexts, motivating the development of automated context engineering. However, existing methods predominantly treat this as a global search problem, seeking a single context strategy that maximizes average performance across a dataset. This restrictive assumption overlooks the fact that different inputs often require distinct guidance, leaving substantial instance-level performance gains untapped. In this paper, we propose a paradigm shift by formulating context engineering as a recommendation problem. We introduce \textbf{Neural Collaborative Context Engineering (NCCE)}, a framework that transitions optimization from a static global search to dynamic, instance-wise routing. NCCE first bootstraps a diverse catalog of anchor contexts and then employs a novel \textbf{Context-CF Co-Evolution} mechanism. This stage establishes a synergistic feedback loop: a lightweight Neural Collaborative Filtering (NCF) model learns instance-context preferences to guide the generation of specialized context variants, while the newly evaluated contexts continuously refine the NCF model's understanding of latent preferences. At inference time, the trained NCF model acts as a context router, dynamically assigning the most suitable context strategy to each unseen instance. Theoretical Proofs and comprehensive experiments demonstrate that by matching individual inputs with their optimal contexts, NCCE significantly improves task accuracy, highlighting the critical importance of personalization in LLM context engineering.

Qingyao Li, Weiwen Liu, Weinan Zhang et al.

Recent advancements in Large Language Models (LLMs) have successfully employed search-based strategies to enhance code generation. However, existing methods typically rely on static, sparse public test cases for verification, leading to pseudo-correctness -- where solutions overfit the visible public tests but fail to generalize to hidden test cases. We argue that optimizing against a fixed, weak environment inherently limits robustness. To address this, we propose AdverMCTS, a novel adversarial Monte Carlo Tree Search framework that combats pseudo-correctness by coupling code search with active vulnerability discovery. AdverMCTS formulates generation as a minimax-style game between a Solver agent, which synthesizes code candidates, and an Attacker agent, which evolves to generate targeted corner test cases that exploit logical divergences in the current code pool. These discovered tests form a dynamic, progressively hostile filter that penalizes fragile reasoning. Extensive experiments demonstrate that AdverMCTS significantly outperforms state-of-the-art baselines, effectively reducing false positive rates and forcing the model to generalize beyond the initial constraints. The resources of this work are available at https://anonymous.4open.science/r/AdverMCTS_open-A255.

Zheng Wu, Heyuan Huang, Yanjia Yang et al.

As multimodal large language models advance rapidly, the automation of mobile tasks has become increasingly feasible through the use of mobile-use agents that mimic human interactions from graphical user interface. To further enhance mobile-use agents, previous studies employ demonstration learning to improve mobile-use agents from human demonstrations. However, these methods focus solely on the explicit intention flows of humans (e.g., step sequences) while neglecting implicit intention flows (e.g., personal preferences), which makes it difficult to construct personalized mobile-use agents. In this work, to evaluate the \textbf{I}ntention \textbf{A}lignment \textbf{R}ate between mobile-use agents and humans, we first collect \textbf{MobileIAR}, a dataset containing human-intent-aligned actions and ground-truth actions. This enables a comprehensive assessment of the agents' understanding of human intent. Then we propose \textbf{IFRAgent}, a framework built upon \textbf{I}ntention \textbf{F}low \textbf{R}ecognition from human demonstrations. IFRAgent analyzes explicit intention flows from human demonstrations to construct a query-level vector library of standard operating procedures (SOP), and analyzes implicit intention flows to build a user-level habit repository. IFRAgent then leverages a SOP extractor combined with retrieval-augmented generation and a query rewriter to generate personalized query and SOP from a raw ambiguous query, enhancing the alignment between mobile-use agents and human intent. Experimental results demonstrate that IFRAgent outperforms baselines by an average of 6.79\% (32.06\% relative improvement) in human intention alignment rate and improves step completion rates by an average of 5.30\% (26.34\% relative improvement). The codes are available at https://github.com/MadeAgents/Quick-on-the-Uptake.

Hao Guan, Lingyue Fu, Shao Zhang et al.

As autonomous code agents move toward end-to-end software development, evaluating their practical autonomy becomes critical. Current benchmarks hide friction by testing agents in pre-configured environments, and their static evaluation pipelines frequently fail when parsing fully autonomous trajectories. We address these limitations with SWE-Cycle, a benchmark of 489 rigorously filtered instances. SWE-Cycle evaluates agents across three isolated tasks, including environment reconstruction, code implementation, and verification test generation, as well as an end-to-end FullCycle task that integrates all three. The FullCycle task requires agents to work autonomously in a bare repository without human scaffolding. To reliably assess these complex execution paths, we developed SWE-Judge. By combining static code review with dynamic testing, this execution-capable evaluation agent accurately verifies functional correctness and eliminates the systematic measurement errors of traditional static parsers. We evaluate code agents powered by six state-of-the-art LLMs across these four tasks. The results reveal a sharp drop in solve rates when transitioning from isolated tasks to FullCycle execution, exposing critical bottlenecks in handling cross-phase dependencies and maintaining code quality. Together, SWE-Cycle and SWE-Judge provide a comprehensive framework for accurately measuring the end-to-end capabilities of autonomous software agents.

Xu Huang, Weiwen Liu, Xiaolong Chen et al.

As Large Language Models (LLMs) have shown significant intelligence, the progress to leverage LLMs as planning modules of autonomous agents has attracted more attention. This survey provides the first systematic view of LLM-based agents planning, covering recent works aiming to improve planning ability. We provide a taxonomy of existing works on LLM-Agent planning, which can be categorized into Task Decomposition, Plan Selection, External Module, Reflection and Memory. Comprehensive analyses are conducted for each direction, and further challenges for the field of research are discussed.

Lingyue Fu, Hao Guan, Bolun Zhang et al.

As Large Language Models (LLMs) demonstrate increasingly sophisticated code processing capabilities, evaluating their performance on engineering-level code remains challenging. Existing repository-level benchmarks primarily focus on single scenarios, such as code generation or bug fixing, without adequately capturing the diversity and complexity of real-world software or project engineering workflows. Furthermore, these benchmarks suffer from limited controllability in question positioning and reliability issues in their generated test cases. To address these limitations, we present CorePipe, a fully automated pipeline that converts repositories into comprehensive test cases, and introduce CoreCodeBench, a configurable multi-scenario repository-level benchmark. To simulate real engineering scenarios, CorePipe generates three types of atomic questions (Development, BugFix, and Test-Driven Development) specifically targeting core code segments. These atomic questions are further combined into three types of composite questions, with difficulty levels flexibly adjusted through hyperparameter tuning. CoreCodeBench provides a comprehensive and extensive repository-level benchmark to investigate the applicability of LLMs in real-world engineering projects. Experiments with 16 LLMs across diverse scenarios reveal varying capabilities and offer multi-dimensional insights into LLM performance in engineering contexts. The code for CorePipe is available at https://github.com/AGI-Eval-Official/CoreCodeBench, and the data for CoreCodeBench can be accessed at https://huggingface.co/collections/tubehhh/corecodebench-68256d2faabf4b1610a08caa.

Zhiyuan Yao, Zishan Xu, Yifu Guo et al.

With the rise of the Agent Web and Model Context Protocol (MCP), the agent ecosystem is evolving into an open collaborative network, exponentially increasing accessible tools. However, current architectures face severe scalability and generality bottlenecks. To address this, we propose ToolACE-MCP, a pipeline for training history-aware routers to empower precise navigation in large-scale ecosystems. By leveraging a dependency-rich candidate Graph to synthesize multi-turn trajectories, we effectively train routers with dynamic context understanding to create the plug-and-play Light Routing Agent. Experiments on the real-world benchmarks MCP-Universe and MCP-Mark demonstrate superior performance. Notably, ToolACE-MCP exhibits critical properties for the future Agent Web: it not only generalizes to multi-agent collaboration with minimal adaptation but also maintains exceptional robustness against noise and scales effectively to massive candidate spaces. These findings provide a strong empirical foundation for universal orchestration in open-ended ecosystems.

Shuai Pan, Yixiang Liu, Jiaye Gao et al.

Large language model (LLM) agent systems are increasingly expected to improve after deployment, but existing work often decouples two adaptation targets: skill evolution and multi-agent system (MAS) restructuring. This separation can create organization bottlenecks, context pressure, and mis-specialization. We present SkillMAS, a non-parametric framework for adaptive specialization in multi-agent systems that couples skill evolution with MAS restructuring. SkillMAS uses Utility Learning to assign credit from verified execution traces, bounded skill evolution to refine reusable procedures without unfiltered library growth, and evidence-gated MAS restructuring when retained failures and Executor Utility indicate a structural mismatch. Across embodied manipulation, command-line execution, and retail workflows, SkillMAS is competitive under the reported harnesses while clarifying how post-deployment specialization is attributed, updated, and applied.

Yuanyi Song, Heyuan Huang, Qiqiang Lin et al.

The rapid advancement of multimodal large language models has enabled agents to operate mobile devices by directly interacting with graphical user interfaces, opening new possibilities for mobile automation. However, real-world mobile tasks are often complex and allow for multiple valid solutions. This contradicts current mobile agent evaluation standards: offline static benchmarks can only validate a single predefined "golden path", while online dynamic testing is constrained by the complexity and non-reproducibility of real devices, making both approaches inadequate for comprehensively assessing agent capabilities. To bridge the gap between offline and online evaluation and enhance testing stability, this paper introduces a novel graph-structured benchmarking framework. By modeling the finite states observed during real-device interactions, it achieves static simulation of dynamic behaviors. Building on this, we develop ColorBench, a benchmark focused on complex long-horizon tasks. It supports evaluation of multiple valid solutions, subtask completion rate statistics, and atomic-level capability analysis. ColorBench contains 175 tasks (74 single-app, 101 cross-app) with an average length of over 13 steps. Each task includes at least two correct paths and several typical error paths, enabling quasi-dynamic interaction. By evaluating ColorBench across various baselines, we discover limitations of existing models and propose improvement directions and feasible technical pathways to enhance agents' performance on complex, long-horizon problems based on experimental results. Code and data are available at: https://github.com/MadeAgents/ColorBench.

Xu Huang, Yuefeng Huang, Weiwen Liu et al.

Tool invocation is a crucial mechanism for extending the capabilities of Large Language Models (LLMs) and has recently garnered significant attention. It enables LLMs to solve complex problems through tool calls while accessing up-to-date world knowledge. However, existing work primarily focuses on the fundamental ability of LLMs to invoke tools for problem-solving, without considering personalized constraints in tool invocation. In this work, we introduce the concept of Personalized Tool Invocation and define two key tasks: Tool Preference and Profile-dependent Query. Tool Preference addresses user preferences when selecting among functionally similar tools, while Profile-dependent Query considers cases where a user query lacks certain tool parameters, requiring the model to infer them from the user profile. To tackle these challenges, we propose PTool, a data synthesis framework designed for personalized tool invocation. Additionally, we construct \textbf{PTBench}, the first benchmark for evaluating personalized tool invocation. We then fine-tune various open-source models, demonstrating the effectiveness of our framework and providing valuable insights. Our benchmark is public at https://github.com/hyfshadow/PTBench.

Fangwen Wu, Zheng Wu, Jihong Wang et al.

With the rapid development of (multimodal) large language model-based agents, the landscape of agentic service management has evolved from single-agent systems to multi-agent systems, and now to massive-agent ecosystems. Current massive-agent ecosystems face growing challenges, including impersonal service experiences, a lack of standardization, and untrustworthy behavior. To address these issues, we propose ColorEcosystem, a novel blueprint designed to enable personalized, standardized, and trustworthy agentic service at scale. Concretely, ColorEcosystem consists of three key components: agent carrier, agent store, and agent audit. The agent carrier provides personalized service experiences by utilizing user-specific data and creating a digital twin, while the agent store serves as a centralized, standardized platform for managing diverse agentic services. The agent audit, based on the supervision of developer and user activities, ensures the integrity and credibility of both service providers and users. Through the analysis of challenges, transitional forms, and practical considerations, the ColorEcosystem is poised to power personalized, standardized, and trustworthy agentic service across massive-agent ecosystems. Meanwhile, we have also implemented part of ColorEcosystem's functionality, and the relevant code is open-sourced at https://github.com/opas-lab/color-ecosystem.

Weiwen Liu, Yunjia Xi, Jiarui Qin et al.

As the final stage of the multi-stage recommender system (MRS), re-ranking directly affects user experience and satisfaction by rearranging the input ranking lists, and thereby plays a critical role in MRS. With the advances in deep learning, neural re-ranking has become a trending topic and been widely applied in industrial applications. This review aims at integrating re-ranking algorithms into a broader picture, and paving ways for more comprehensive solutions for future research. For this purpose, we first present a taxonomy of current methods on neural re-ranking. Then we give a description of these methods along with the historic development according to their objectives. The network structure, personalization, and complexity are also discussed and compared. Next, we provide benchmarks of the major neural re-ranking models and quantitatively analyze their re-ranking performance. Finally, the review concludes with a discussion on future prospects of this field. A list of papers discussed in this review, the benchmark datasets, our re-ranking library LibRerank, and detailed parameter settings are publicly available at https://github.com/LibRerank-Community/LibRerank.

Yun He, Yin Zhang, Weiwen Liu et al.

User-generated item lists are popular on many platforms. Examples include video-based playlists on YouTube, image-based lists (or"boards") on Pinterest, book-based lists on Goodreads, and answer-based lists on question-answer forums like Zhihu. As users create these lists, a common challenge is in identifying what items to curate next. Some lists are organized around particular genres or topics, while others are seemingly incoherent, reflecting individual preferences for what items belong together. Furthermore, this heterogeneity in item consistency may vary from platform to platform, and from sub-community to sub-community. Hence, this paper proposes a generalizable approach for user-generated item list continuation. Complementary to methods that exploit specific content patterns (e.g., as in song-based playlists that rely on audio features), the proposed approach models the consistency of item lists based on human curation patterns, and so can be deployed across a wide range of varying item types (e.g., videos, images, books). A key contribution is in intelligently combining two preference models via a novel consistency-aware gating network - a general user preference model that captures a user's overall interests, and a current preference priority model that captures a user's current (as of the most recent item) interests. In this way, the proposed consistency-aware recommender can dynamically adapt as user preferences evolve. Evaluation over four datasets(of songs, books, and answers) confirms these observations and demonstrates the effectiveness of the proposed model versus state-of-the-art alternatives. Further, all code and data are available at https://github.com/heyunh2015/ListContinuation_WSDM2020.

Chenyu Zhou, Huacan Chai, Wenteng Chen et al.

Large language model (LLM) agents are increasingly built less by changing model weights than by reorganizing the runtime around them. Capabilities that earlier systems expected the model to recover internally are now externalized into memory stores, reusable skills, interaction protocols, and the surrounding harness that makes these modules reliable in practice. This paper reviews that shift through the lens of externalization. Drawing on the idea of cognitive artifacts, we argue that agent infrastructure matters not merely because it adds auxiliary components, but because it transforms hard cognitive burdens into forms that the model can solve more reliably. Under this view, memory externalizes state across time, skills externalize procedural expertise, protocols externalize interaction structure, and harness engineering serves as the unification layer that coordinates them into governed execution. We trace a historical progression from weights to context to harness, analyze memory, skills, and protocols as three distinct but coupled forms of externalization, and examine how they interact inside a larger agent system. We further discuss the trade-off between parametric and externalized capability, identify emerging directions such as self-evolving harnesses and shared agent infrastructure, and discuss open challenges in evaluation, governance, and the long-term co-evolution of models and external infrastructure. The result is a systems-level framework for explaining why practical agent progress increasingly depends not only on stronger models, but on better external cognitive infrastructure.

Shuai Wang, Weiwen Liu, Jingxuan Chen et al.

Recent advances in foundation models, particularly Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs), have facilitated the development of intelligent agents capable of performing complex tasks. By leveraging the ability of (M)LLMs to process and interpret Graphical User Interfaces (GUIs), these agents can autonomously execute user instructions, simulating human-like interactions such as clicking and typing. This survey consolidates recent research on (M)LLM-based GUI agents, highlighting key innovations in data resources, frameworks, and applications. We begin by reviewing representative datasets and benchmarks, followed by an overview of a generalized, unified framework that encapsulates the essential components of prior studies, supported by a detailed taxonomy. Additionally, we explore relevant commercial applications. Drawing insights from existing work, we identify key challenges and propose future research directions. We hope this survey will inspire further advancements in the field of (M)LLM-based GUI agents.

Yingxuan Yang, Huacan Chai, Yuanyi Song et al.

The rapid development of large language models (LLMs) has led to the widespread deployment of LLM agents across diverse industries, including customer service, content generation, data analysis, and even healthcare. However, as more LLM agents are deployed, a major issue has emerged: there is no standard way for these agents to communicate with external tools or data sources. This lack of standardized protocols makes it difficult for agents to work together or scale effectively, and it limits their ability to tackle complex, real-world tasks. A unified communication protocol for LLM agents could change this. It would allow agents and tools to interact more smoothly, encourage collaboration, and triggering the formation of collective intelligence. In this paper, we provide the first comprehensive analysis of existing agent protocols, proposing a systematic two-dimensional classification that differentiates context-oriented versus inter-agent protocols and general-purpose versus domain-specific protocols. Additionally, we conduct a comparative performance analysis of these protocols across key dimensions such as security, scalability, and latency. Finally, we explore the future landscape of agent protocols by identifying critical research directions and characteristics necessary for next-generation protocols. These characteristics include adaptability, privacy preservation, and group-based interaction, as well as trends toward layered architectures and collective intelligence infrastructures. We expect this work to serve as a practical reference for both researchers and engineers seeking to design, evaluate, or integrate robust communication infrastructures for intelligent agents.

Jingxuan Chen, Derek Yuen, Bin Xie et al.

Smartphone agents are increasingly important for helping users control devices efficiently, with (Multimodal) Large Language Model (MLLM)-based approaches emerging as key contenders. Fairly comparing these agents is essential but challenging, requiring a varied task scope, the integration of agents with different implementations, and a generalisable evaluation pipeline to assess their strengths and weaknesses. In this paper, we present SPA-Bench, a comprehensive SmartPhone Agent Benchmark designed to evaluate (M)LLM-based agents in an interactive environment that simulates real-world conditions. SPA-Bench offers three key contributions: (1) A diverse set of tasks covering system and third-party apps in both English and Chinese, focusing on features commonly used in daily routines; (2) A plug-and-play framework enabling real-time agent interaction with Android devices, integrating over ten agents with the flexibility to add more; (3) A novel evaluation pipeline that automatically assesses agent performance across multiple dimensions, encompassing seven metrics related to task completion and resource consumption. Our extensive experiments across tasks and agents reveal challenges like interpreting mobile user interfaces, action grounding, memory retention, and execution costs. We propose future research directions to ease these difficulties, moving closer to real-world smartphone agent applications. SPA-Bench is available at https://ai-agents-2030.github.io/SPA-Bench/.

Chen Chen, Xinlong Hao, Weiwen Liu et al.

Large Language Models (LLMs) have demonstrated significant potential in decision-making and reasoning, particularly when integrated with various tools to effectively solve complex problems. However, existing benchmarks for evaluating LLMs' tool usage face several limitations: (1) limited evaluation scenarios, often lacking assessments in real multi-turn dialogue contexts; (2) narrow evaluation dimensions, with insufficient detailed assessments of how LLMs use tools; and (3) reliance on LLMs or real API executions for evaluation, which introduces significant overhead. To address these challenges, we introduce ACEBench, a comprehensive benchmark for assessing tool usage in LLMs. ACEBench categorizes data into three primary types based on evaluation methodology: Normal, Special, and Agent. "Normal" evaluates tool usage in basic scenarios; "Special" evaluates tool usage in situations with ambiguous or incomplete instructions; "Agent" evaluates tool usage through multi-agent interactions to simulate real-world, multi-turn dialogues. We conducted extensive experiments using ACEBench, analyzing various LLMs in-depth and providing a more granular examination of error causes across different data types.

Yichun Yin, Wenyong Huang, Kaikai Song et al.

We present Pangu Ultra, a Large Language Model (LLM) with 135 billion parameters and dense Transformer modules trained on Ascend Neural Processing Units (NPUs). Although the field of LLM has been witnessing unprecedented advances in pushing the scale and capability of LLM in recent years, training such a large-scale model still involves significant optimization and system challenges. To stabilize the training process, we propose depth-scaled sandwich normalization, which effectively eliminates loss spikes during the training process of deep models. We pre-train our model on 13.2 trillion diverse and high-quality tokens and further enhance its reasoning capabilities during post-training. To perform such large-scale training efficiently, we utilize 8,192 Ascend NPUs with a series of system optimizations. Evaluations on multiple diverse benchmarks indicate that Pangu Ultra significantly advances the state-of-the-art capabilities of dense LLMs such as Llama 405B and Mistral Large 2, and even achieves competitive results with DeepSeek-R1, whose sparse model structure contains much more parameters. Our exploration demonstrates that Ascend NPUs are capable of efficiently and effectively training dense models with more than 100 billion parameters. Our model and system will be available for our commercial customers.