Large language model (LLM) agents have evolved to intelligently process information, make decisions, and interact with users or tools. A key capability is the integration of long-term memory capabilities, enabling these agents to draw upon historical interactions and knowledge. However, the growing memory size and need for semantic structuring pose significant challenges. In this work, we propose an autonomous memory augmentation approach, MemInsight, to enhance semantic data representation and retrieval mechanisms. By leveraging autonomous augmentation to historical interactions, LLM agents are shown to deliver more accurate and contextualized responses. We empirically validate the efficacy of our proposed approach in three task scenarios; conversational recommendation, question answering and event summarization. On the LLM-REDIAL dataset, MemInsight boosts persuasiveness of recommendations by up to 14%. Moreover, it outperforms a RAG baseline by 34% in recall for LoCoMo retrieval. Our empirical results show the potential of MemInsight to enhance the contextual performance of LLM agents across multiple tasks.
Ensuring the reliability and effectiveness of software release decisions is critical, particularly in safety-critical domains like automotive systems. Precise analysis of release validation data, often presented in tabular form, plays a pivotal role in this process. However, traditional methods that rely on manual analysis of extensive test datasets and validation metrics are prone to delays and high costs. Large Language Models (LLMs) offer a promising alternative but face challenges in analytical reasoning, contextual understanding, handling out-of-scope queries, and processing structured test data consistently; limitations that hinder their direct application in safety-critical scenarios. This paper introduces GateLens, an LLM-based tool for analyzing tabular data in the automotive domain. GateLens translates natural language queries into Relational Algebra (RA) expressions and then generates optimized Python code. It outperforms the baseline system on benchmarking datasets, achieving higher F1 scores and handling complex and ambiguous queries with greater robustness. Ablation studies confirm the critical role of the RA module, with performance dropping sharply when omitted. Industrial evaluations reveal that GateLens reduces analysis time by over 80% while maintaining high accuracy and reliability. As demonstrated by presented results, GateLens achieved high performance without relying on few-shot examples, showcasing strong generalization across various query types from diverse company roles. Insights from deploying GateLens with a partner automotive company offer practical guidance for integrating AI into critical workflows such as release validation. Results show that by automating test result analysis, GateLens enables faster, more informed, and dependable release decisions, and can thus advance software scalability and reliability in automotive systems.
Alignment of Large Language models (LLMs) is crucial for safe and trustworthy deployment in applications. Reinforcement learning from human feedback (RLHF) has emerged as an effective technique to align LLMs to human preferences and broader utilities, but it requires updating billions of model parameters, which is computationally expensive. Controlled Decoding, by contrast, provides a mechanism for aligning a model at inference time without retraining. However, single-agent decoding approaches often struggle to adapt to diverse tasks due to the complexity and variability inherent in these tasks. To strengthen the test-time performance w.r.t the target task, we propose a mixture of agent-based decoding strategies leveraging the existing off-the-shelf aligned LLM policies. Treating each prior policy as an agent in the spirit of mixture of agent collaboration, we develop a decoding method that allows for inference-time alignment through a token-level selection strategy among multiple agents. For each token, the most suitable LLM is dynamically chosen from a pool of models based on a long-term utility metric. This policy-switching mechanism ensures optimal model selection at each step, enabling efficient collaboration and alignment among LLMs during decoding. Theoretical analysis of our proposed algorithm establishes optimal performance with respect to the target task represented via a target reward for the given off-the-shelf models. We conduct comprehensive empirical evaluations with open-source aligned models on diverse tasks and preferences, which demonstrates the merits of this approach over single-agent decoding baselines. Notably, Collab surpasses the current SoTA decoding strategy, achieving an improvement of up to 1.56x in average reward and 71.89% in GPT-4 based win-tie rate.
The recent DeepSeek-R1 has showcased the emergence of reasoning capabilities in LLMs through reinforcement learning (RL) with rule-based rewards. Building on this idea, we are the first to explore how rule-based RL can enhance the reasoning capabilities of multimodal large language models (MLLMs) for graphic user interface (GUI) action prediction tasks. To this end, we curate a small yet high-quality dataset of 136 challenging tasks, encompassing five common action types on mobile devices. We also introduce a unified rule-based action reward, enabling model optimization via policy-based algorithms such as Group Relative Policy Optimization (GRPO). Experimental results demonstrate that our proposed data-efficient model, UI-R1-3B, achieves substantial improvements on both in-domain (ID) and out-of-domain (OOD) tasks. Specifically, on the ID benchmark AndroidControl, the action type accuracy improves by 15%, while grounding accuracy increases by 10.3%, compared with the base model (i.e. Qwen2.5-VL-3B). On the OOD GUI grounding benchmark ScreenSpot-Pro, our model surpasses the base model by 6.0% and achieves competitive performance with larger models (e.g., OS-Atlas-7B), which are trained via supervised fine-tuning (SFT) on 76K data. These results underscore the potential of rule-based reinforcement learning to advance GUI understanding and control, paving the way for future research in this domain.
Successful agent-human partnerships require that any agent generated information is understandable to the human, and that the human can easily steer the agent towards a goal. Such effective communication requires the agent to develop a finer-level notion of what is understandable to the human. State-of-the-art agents, including LLMs, lack this detailed notion of understandability because they only capture average human sensibilities from the training data, and therefore afford limited steerability (e.g., requiring non-trivial prompt engineering). In this paper, instead of only relying on data, we argue for developing generalizable, domain-agnostic measures of understandability that can be used as directives for these agents. Existing research on understandability measures is fragmented, we survey various such efforts across domains, and lay a cognitive-science-rooted groundwork for more coherent and domain-agnostic research investigations in future.
Large Language Models (LLMs) are increasingly relied upon for coding tasks, yet in most scenarios it is assumed that all relevant information can be either accessed in context or matches their training data. We posit that LLMs can benefit from the ability to interactively explore a codebase to gather the information relevant to their task. To achieve this, we present a textual environment, namely debug-gym, for developing LLM-based agents in an interactive coding setting. Our environment is lightweight and provides a preset of useful tools, such as a Python debugger (pdb), designed to facilitate an LLM-based agent's interactive debugging. Beyond coding and debugging tasks, this approach can be generalized to other tasks that would benefit from information-seeking behavior by an LLM agent.
The era of intelligent agents is upon us, driven by revolutionary advancements in large language models. Large Language Model (LLM) agents, with goal-driven behaviors and dynamic adaptation capabilities, potentially represent a critical pathway toward artificial general intelligence. This survey systematically deconstructs LLM agent systems through a methodology-centered taxonomy, linking architectural foundations, collaboration mechanisms, and evolutionary pathways. We unify fragmented research threads by revealing fundamental connections between agent design principles and their emergent behaviors in complex environments. Our work provides a unified architectural perspective, examining how agents are constructed, how they collaborate, and how they evolve over time, while also addressing evaluation methodologies, tool applications, practical challenges, and diverse application domains. By surveying the latest developments in this rapidly evolving field, we offer researchers a structured taxonomy for understanding LLM agents and identify promising directions for future research. The collection is available at https://github.com/luo-junyu/Awesome-Agent-Papers.
Quantitative investment (quant) is an emerging, technology-driven approach in asset management, increasingy shaped by advancements in artificial intelligence. Recent advances in deep learning and large language models (LLMs) for quant finance have improved predictive modeling and enabled agent-based automation, suggesting a potential paradigm shift in this field. In this survey, taking alpha strategy as a representative example, we explore how AI contributes to the quantitative investment pipeline. We first examine the early stage of quant research, centered on human-crafted features and traditional statistical models with an established alpha pipeline. We then discuss the rise of deep learning, which enabled scalable modeling across the entire pipeline from data processing to order execution. Building on this, we highlight the emerging role of LLMs in extending AI beyond prediction, empowering autonomous agents to process unstructured data, generate alphas, and support self-iterative workflows.
Adapting Large Language Models (LLMs) to downstream tasks using Reinforcement Learning (RL) has proven to be an effective approach. However, LLMs do not inherently define the structure of an agent for RL training, particularly in terms of defining the action space. This paper studies learning a compact latent action space to enhance the controllability and exploration of RL for LLMs. We propose Controlling Large Language Models with Latent Actions (CoLA), a framework that integrates a latent action space into pre-trained LLMs. We apply CoLA to the Llama-3.1-8B model. Our experiments demonstrate that, compared to RL with token-level actions, CoLA's latent action enables greater semantic diversity in text generation. For enhancing downstream tasks, we show that CoLA with RL achieves a score of 42.4 on the math500 benchmark, surpassing the baseline score of 38.2, and reaches 68.2 when augmented with a Monte Carlo Tree Search variant. Furthermore, CoLA with RL consistently improves performance on agent-based tasks without degrading the pre-trained LLM's capabilities, unlike the baseline. Finally, CoLA reduces computation time by half in tasks involving enhanced thinking prompts for LLMs by RL. These results highlight CoLA's potential to advance RL-based adaptation of LLMs for downstream applications.
Current AI counseling systems struggle with maintaining effective long-term client engagement. Through formative research with counselors and a systematic literature review, we identified five key design considerations for AI counseling interactions. Based on these insights, we propose CA+, a Cognition Augmented counselor framework enhancing contextual understanding through three components: (1) Therapy Strategies Module: Implements hierarchical Goals-Session-Action planning with bidirectional adaptation based on client feedback; (2) Communication Form Module: Orchestrates parallel guidance and empathy pathways for balanced therapeutic progress and emotional resonance; (3) Information Management: Utilizes client profile and therapeutic knowledge databases for dynamic, context-aware interventions. A three-day longitudinal study with 24 clients demonstrates CA+'s significant improvements in client engagement, perceived empathy, and overall satisfaction compared to a baseline system. Besides, two licensed counselors confirm its high professionalism. Our research demonstrates the potential for enhancing LLM engagement in psychological counseling dialogues through cognitive theory, which may inspire further innovations in computational interaction in the future.
Advancements in retrieving accessible information have evolved faster in the last few years compared to the decades since the internet's creation. Search engines, like Google, have been the number one way to find relevant data. They have always relied on the user's abilities to find the best information in its billions of links and sources at everybody's fingertips. The advent of large language models (LLMs) has completely transformed the field of information retrieval. The LLMs excel not only at retrieving relevant knowledge but also at summarizing it effectively, making information more accessible and consumable for users. On top of it, the rise of AI Agents has introduced another aspect to information retrieval i.e. dynamic information retrieval which enables the integration of real-time data such as weather forecasts, and financial data with the knowledge base to curate context-aware knowledge. However, despite these advancements the agents remain susceptible to issues of bias and fairness, challenges deeply rooted within the knowledge base and training of LLMs. This study introduces a novel approach to bias-aware knowledge retrieval by leveraging agentic framework and the innovative use of bias detectors as tools to identify and highlight inherent biases in the retrieved content. By empowering users with transparency and awareness, this approach aims to foster more equitable information systems and promote the development of responsible AI.
Generative retrieval (GR) has revolutionized document retrieval with the advent of large language models (LLMs), and LLM-based GR is gradually being adopted by the industry. Despite its remarkable advantages and potential, LLM-based GR suffers from hallucination and generates documents that are irrelevant to the query in some instances, severely challenging its credibility in practical applications. We thereby propose an optimized GR framework designed to alleviate retrieval hallucination, which integrates knowledge distillation reasoning in model training and incorporate decision agent to further improve retrieval precision. Specifically, we employ LLMs to assess and reason GR retrieved query-document (q-d) pairs, and then distill the reasoning data as transferred knowledge to the GR model. Moreover, we utilize a decision agent as post-processing to extend the GR retrieved documents through retrieval model and select the most relevant ones from multi perspectives as the final generative retrieval result. Extensive offline experiments on real-world datasets and online A/B tests on Fund Search and Insurance Search in Alipay demonstrate our framework's superiority and effectiveness in improving search quality and conversion gains.
While large language model (LLM)-based chatbots have been applied for effective engagement in credit dialogues, their capacity for dynamic emotional expression remains limited. Current agents primarily rely on passive empathy rather than affective reasoning. For instance, when faced with persistent client negativity, the agent should employ strategic emotional adaptation by expressing measured anger to discourage counterproductive behavior and guide the conversation toward resolution. This context-aware emotional modulation is essential for imitating the nuanced decision-making of human negotiators. This paper introduces an EQ-negotiator that combines emotion sensing from pre-trained language models (PLMs) with emotional reasoning based on Game Theory and Hidden Markov Models. It takes into account both the current and historical emotions of the client to better manage and address negative emotions during interactions. By fine-tuning pre-trained language models (PLMs) on public emotion datasets and validating them on the credit dialogue datasets, our approach enables LLM-based agents to effectively capture shifts in client emotions and dynamically adjust their response tone based on our emotion decision policies in real-world financial negotiations. This EQ-negotiator can also help credit agencies foster positive client relationships, enhancing satisfaction in credit services.
Reasoning segmentation (RS) aims to identify and segment objects of interest based on implicit text queries. As such, RS is a catalyst for embodied AI agents, enabling them to interpret high-level commands without requiring explicit step-by-step guidance. However, current RS approaches rely heavily on the visual perception capabilities of multimodal large language models (LLMs), leading to several major limitations. First, they struggle with queries that require multiple steps of reasoning or those that involve complex spatial/temporal relationships. Second, they necessitate LLM fine-tuning, which may require frequent updates to maintain compatibility with contemporary LLMs and may increase risks of catastrophic forgetting during fine-tuning. Finally, being primarily designed for static images or offline video processing, they scale poorly to online video data. To address these limitations, we propose an agent framework that disentangles perception and reasoning for online video RS without LLM fine-tuning. Our innovation is the introduction of a just-in-time digital twin concept, where -- given an implicit query -- a LLM plans the construction of a low-level scene representation from high-level video using specialist vision models. We refer to this approach to creating a digital twin as "just-in-time" because the LLM planner will anticipate the need for specific information and only request this limited subset instead of always evaluating every specialist model. The LLM then performs reasoning on this digital twin representation to identify target objects. To evaluate our approach, we introduce a new comprehensive video reasoning segmentation benchmark comprising 200 videos with 895 implicit text queries. The benchmark spans three reasoning categories (semantic, spatial, and temporal) with three different reasoning chain complexity.
Analyzing operating room (OR) workflows to derive quantitative insights into OR efficiency is important for hospitals to maximize patient care and financial sustainability. Prior work on OR-level workflow analysis has relied on end-to-end deep neural networks. While these approaches work well in constrained settings, they are limited to the conditions specified at development time and do not offer the flexibility necessary to accommodate the OR workflow analysis needs of various OR scenarios (e.g., large academic center vs. rural provider) without data collection, annotation, and retraining. Reasoning segmentation (RS) based on foundation models offers this flexibility by enabling automated analysis of OR workflows from OR video feeds given only an implicit text query related to the objects of interest. Due to the reliance on large language model (LLM) fine-tuning, current RS approaches struggle with reasoning about semantic/spatial relationships and show limited generalization to OR video due to variations in visual characteristics and domain-specific terminology. To address these limitations, we first propose a novel digital twin (DT) representation that preserves both semantic and spatial relationships between the various OR components. Then, building on this foundation, we propose ORDiRS (Operating Room Digital twin representation for Reasoning Segmentation), an LLM-tuning-free RS framework that reformulates RS into a "reason-retrieval-synthesize" paradigm. Finally, we present ORDiRS-Agent, an LLM-based agent that decomposes OR workflow analysis queries into manageable RS sub-queries and generates responses by combining detailed textual explanations with supporting visual evidence from RS. Experimental results on both an in-house and a public OR dataset demonstrate that our ORDiRS achieves a cIoU improvement of 6.12%-9.74% compared to the existing state-of-the-arts.
We present an agentic framework, Thinker, which achieves state of art performance in challenging reasoning tasks for realistic customer service scenarios that involve complex business logic and human interactions via long horizons. On the $\tau$-bench retail dataset, Thinker achieves 82.6\% success rate with GPT-4o (version 2024-06-01) (baseline: 68.3\%), and 81.9\% success rate with Llama-3.1 405B (baseline: 49.6\%), without any fine-tuning. Thinker effectively closes the gap in reasoning capabilities between the base models by introducing proper structure. The key features of the Thinker framework are: (1) State-Machine Augmented Generation (SMAG), which represents business logic as state machines and the LLM uses state machines as tools. (2) Delegation of tasks from the main reasoning loop to LLM-powered tools. (3) Adaptive context management. Our prompting-only solution achieves signficant gains, while still maintaining a standard agentic architecture with a ReAct style reasoning loop. The key is to innovate on the tool interface design, as exemplified by SMAG and the LLM-powered tools.
Recent advancements in 2D and multimodal models have achieved remarkable success by leveraging large-scale training on extensive datasets. However, extending these achievements to enable free-form interactions and high-level semantic operations with complex 3D/4D scenes remains challenging. This difficulty stems from the limited availability of large-scale, annotated 3D/4D or multi-view datasets, which are crucial for generalizable vision and language tasks such as open-vocabulary and prompt-based segmentation, language-guided editing, and visual question answering (VQA). In this paper, we introduce Feature4X, a universal framework designed to extend any functionality from 2D vision foundation model into the 4D realm, using only monocular video input, which is widely available from user-generated content. The "X" in Feature4X represents its versatility, enabling any task through adaptable, model-conditioned 4D feature field distillation. At the core of our framework is a dynamic optimization strategy that unifies multiple model capabilities into a single representation. Additionally, to the best of our knowledge, Feature4X is the first method to distill and lift the features of video foundation models (e.g. SAM2, InternVideo2) into an explicit 4D feature field using Gaussian Splatting. Our experiments showcase novel view segment anything, geometric and appearance scene editing, and free-form VQA across all time steps, empowered by LLMs in feedback loops. These advancements broaden the scope of agentic AI applications by providing a foundation for scalable, contextually and spatiotemporally aware systems capable of immersive dynamic 4D scene interaction.
Recent research shows that LLMs can simulate ``believable'' human behaviors to power LLM agents via prompt-only methods. In this work, we focus on evaluating and improving LLM's objective ``accuracy'' rather than the subjective ``believability'' in the web action generation task, leveraging a large-scale, real-world dataset collected from online shopping human actions. We present the first comprehensive quantitative evaluation of state-of-the-art LLMs (e.g., DeepSeek-R1, Llama, and Claude) on the task of web action generation. Our results show that fine-tuning LLMs on real-world behavioral data substantially improves their ability to generate actions compared to prompt-only methods. Furthermore, incorporating synthesized reasoning traces into model training leads to additional performance gains, demonstrating the value of explicit rationale in behavior modeling. This work establishes a new benchmark for evaluating LLMs in behavior simulation and offers actionable insights into how real-world action data and reasoning augmentation can enhance the fidelity of LLM agents.
Thematic analysis (TA) is a widely used qualitative approach for uncovering latent meanings in unstructured text data. TA provides valuable insights in healthcare but is resource-intensive. Large Language Models (LLMs) have been introduced to perform TA, yet their applications in healthcare remain unexplored. Here, we propose TAMA: A Human-AI Collaborative Thematic Analysis framework using Multi-Agent LLMs for clinical interviews. We leverage the scalability and coherence of multi-agent systems through structured conversations between agents and coordinate the expertise of cardiac experts in TA. Using interview transcripts from parents of children with Anomalous Aortic Origin of a Coronary Artery (AAOCA), a rare congenital heart disease, we demonstrate that TAMA outperforms existing LLM-assisted TA approaches, achieving higher thematic hit rate, coverage, and distinctiveness. TAMA demonstrates strong potential for automated TA in clinical settings by leveraging multi-agent LLM systems with human-in-the-loop integration by enhancing quality while significantly reducing manual workload.
Prompt engineering has emerged as a powerful technique for guiding large language models (LLMs) toward desired responses, significantly enhancing their performance across diverse tasks. Beyond their role as static predictors, LLMs increasingly function as intelligent agents, capable of reasoning, decision-making, and adapting dynamically to complex environments. However, the theoretical underpinnings of prompt engineering remain largely unexplored. In this paper, we introduce a formal framework demonstrating that transformer models, when provided with carefully designed prompts, can act as a configurable computational system by emulating a ``virtual'' neural network during inference. Specifically, input prompts effectively translate into the corresponding network configuration, enabling LLMs to adjust their internal computations dynamically. Building on this construction, we establish an approximation theory for $\beta$-times differentiable functions, proving that transformers can approximate such functions with arbitrary precision when guided by appropriately structured prompts. Moreover, our framework provides theoretical justification for several empirically successful prompt engineering techniques, including the use of longer, structured prompts, filtering irrelevant information, enhancing prompt token diversity, and leveraging multi-agent interactions. By framing LLMs as adaptable agents rather than static models, our findings underscore their potential for autonomous reasoning and problem-solving, paving the way for more robust and theoretically grounded advancements in prompt engineering and AI agent design.
Architecture design is a critical step in software development. However, creating a high-quality architecture is often costly due to the significant need for human expertise and manual effort. Recently, agents built upon Large Language Models (LLMs) have achieved remarkable success in various software engineering tasks. Despite this progress, the use of agents to automate the architecture design process remains largely unexplored. To address this gap, we envision a Knowledge-based Multi-Agent Architecture Design (MAAD) framework. MAAD uses agents to simulate human roles in the traditional software architecture design process, thereby automating the design process. To empower these agents, MAAD incorporates knowledge extracted from three key sources: 1) existing system designs, 2) authoritative literature, and 3) architecture experts. By envisioning the MAAD framework, we aim to advance the full automation of application-level system development.
This study investigates the elicitation of empathy toward a third party through interaction with social agents. Participants engaged with either a physical robot or a voice-enabled chatbot, both driven by a large language model (LLM) programmed to exhibit either an empathetic tone or remain neutral. The interaction is focused on a fictional character, Katie Banks, who is in a challenging situation and in need of financial donations. The willingness to help Katie, measured by the number of hours participants were willing to volunteer, along with their perceptions of the agent, were assessed for 60 participants. Results indicate that neither robotic embodiment nor empathetic tone significantly influenced participants' willingness to volunteer. While the LLM effectively simulated human empathy, fostering genuine empathetic responses in participants proved challenging.
Large Language Models (LLMs) are increasingly deployed as computer-use agents, autonomously performing tasks within real desktop or web environments. While this evolution greatly expands practical use cases for humans, it also creates serious security exposures. We present SUDO (Screen-based Universal Detox2Tox Offense), a novel attack framework that systematically bypasses refusal trained safeguards in commercial computer-use agents, such as Claude Computer Use. The core mechanism, Detox2Tox, transforms harmful requests (that agents initially reject) into seemingly benign requests via detoxification, secures detailed instructions from advanced vision language models (VLMs), and then reintroduces malicious content via toxification just before execution. Unlike conventional jailbreaks, SUDO iteratively refines its attacks based on a built-in refusal feedback, making it increasingly effective against robust policy filters. In extensive tests spanning 50 real-world tasks and multiple state-of-the-art VLMs, SUDO achieves a stark attack success rate of 24% (with no refinement), and up to 41% (by its iterative refinement) in Claude Computer Use. By revealing these vulnerabilities and demonstrating the ease with which they can be exploited in real-world computing environments, this paper highlights an immediate need for robust, context-aware safeguards. WARNING: This paper includes harmful or offensive model outputs.
We introduce Open Deep Search (ODS) to close the increasing gap between the proprietary search AI solutions, such as Perplexity's Sonar Reasoning Pro and OpenAI's GPT-4o Search Preview, and their open-source counterparts. The main innovation introduced in ODS is to augment the reasoning capabilities of the latest open-source LLMs with reasoning agents that can judiciously use web search tools to answer queries. Concretely, ODS consists of two components that work with a base LLM chosen by the user: Open Search Tool and Open Reasoning Agent. Open Reasoning Agent interprets the given task and completes it by orchestrating a sequence of actions that includes calling tools, one of which is the Open Search Tool. Open Search Tool is a novel web search tool that outperforms proprietary counterparts. Together with powerful open-source reasoning LLMs, such as DeepSeek-R1, ODS nearly matches and sometimes surpasses the existing state-of-the-art baselines on two benchmarks: SimpleQA and FRAMES. For example, on the FRAMES evaluation benchmark, ODS improves the best existing baseline of the recently released GPT-4o Search Preview by 9.7% in accuracy. ODS is a general framework for seamlessly augmenting any LLMs -- for example, DeepSeek-R1 that achieves 82.4% on SimpleQA and 30.1% on FRAMES -- with search and reasoning capabilities to achieve state-of-the-art performance: 88.3% on SimpleQA and 75.3% on FRAMES.
Recent advancements in LLMs have revolutionized motion generation models in embodied applications. While LLM-type auto-regressive motion generation models benefit from training scalability, there remains a discrepancy between their token prediction objectives and human preferences. As a result, models pre-trained solely with token-prediction objectives often generate behaviors that deviate from what humans would prefer, making post-training preference alignment crucial for producing human-preferred motions. Unfortunately, post-training alignment requires extensive preference rankings of motions generated by the pre-trained model, which are costly to annotate, especially in multi-agent settings. Recently, there has been growing interest in leveraging pre-training demonstrations to scalably generate preference data for post-training alignment. However, these methods often adopt an adversarial assumption, treating all pre-trained model-generated samples as unpreferred examples. This adversarial approach overlooks the valuable signal provided by preference rankings among the model's own generations, ultimately reducing alignment effectiveness and potentially leading to misaligned behaviors. In this work, instead of treating all generated samples as equally bad, we leverage implicit preferences encoded in pre-training demonstrations to construct preference rankings among the pre-trained model's generations, offering more nuanced preference alignment guidance with zero human cost. We apply our approach to large-scale traffic simulation and demonstrate its effectiveness in improving the realism of pre-trained model's generated behaviors, making a lightweight 1M motion generation model comparable to SOTA large imitation-based models by relying solely on implicit feedback from pre-training demonstrations, without additional post-training human preference annotations or high computational costs.
Reproducing game bugs, in our case crash bugs in continuously evolving games like Minecraft, is a notoriously manual, time-consuming, and challenging process to automate. Despite the success of LLM-driven bug reproduction in other software domains, games, with their complex interactive environments, remain largely unaddressed. This paper introduces BugCraft, a novel end-to-end framework designed to automate the reproduction of crash bugs in Minecraft directly from user-submitted bug reports, addressing the critical gap in automated game bug reproduction. BugCraft employs a two-stage approach: first, a Step Synthesizer leverages LLMs and Minecraft Wiki knowledge to transform bug reports into high-quality, structured steps to reproduce (S2R). Second, an Action Model, powered by a vision-based LLM agent (GPT-4o) and a custom macro API, executes these S2R steps within Minecraft to trigger the reported crash. To facilitate evaluation, we introduce BugCraft-Bench, a curated dataset of Minecraft crash bug reports. Evaluated on BugCraft-Bench, our framework successfully reproduced 30.23% of crash bugs end-to-end. The Step Synthesizer demonstrated a 66.28% accuracy in generating correct bug reproduction plans, highlighting its effectiveness in interpreting and structuring bug report information. BugCraft demonstrates the feasibility of automated reproduction of crash bugs in complex game environments using LLMs, opening promising avenues for game testing and development. The framework and the BugCraft-Bench dataset pave the way for future research in automated game bug analysis and hold potential for generalization to other interactive game platforms. Finally, we make our code open at https://bugcraft2025.github.io/
The integration of Large Language Models (LLMs) with specialized tools presents new opportunities for intelligent automation systems. However, orchestrating multiple LLM-driven agents to tackle complex tasks remains challenging due to coordination difficulties, inefficient resource utilization, and inconsistent information flow. We present OmniNova, a modular multi-agent automation framework that combines language models with specialized tools such as web search, crawling, and code execution capabilities. OmniNova introduces three key innovations: (1) a hierarchical multi-agent architecture with distinct coordinator, planner, supervisor, and specialist agents; (2) a dynamic task routing mechanism that optimizes agent deployment based on task complexity; and (3) a multi-layered LLM integration system that allocates appropriate models to different cognitive requirements. Our evaluations across 50 complex tasks in research, data analysis, and web interaction domains demonstrate that OmniNova outperforms existing frameworks in task completion rate (87\% vs. baseline 62\%), efficiency (41\% reduced token usage), and result quality (human evaluation score of 4.2/5 vs. baseline 3.1/5). We contribute both a theoretical framework for multi-agent system design and an open-source implementation that advances the state-of-the-art in LLM-based automation systems.
Information retrieval in hour-long videos presents a significant challenge, even for state-of-the-art Vision-Language Models (VLMs), particularly when the desired information is localized within a small subset of frames. Long video data presents challenges for VLMs due to context window limitations and the difficulty of pinpointing frames containing the answer. Our novel video agent, FALCONEye, combines a VLM and a Large Language Model (LLM) to search relevant information along the video, and locate the frames with the answer. FALCONEye novelty relies on 1) the proposed meta-architecture, which is better suited to tackle hour-long videos compared to short video approaches in the state-of-the-art; 2) a new efficient exploration algorithm to locate the information using short clips, captions and answer confidence; and 3) our state-of-the-art VLMs calibration analysis for the answer confidence. Our agent is built over a small-size VLM and a medium-size LLM being accessible to run on standard computational resources. We also release FALCON-Bench, a benchmark to evaluate long (average > 1 hour) Video Answer Search challenges, highlighting the need for open-ended question evaluation. Our experiments show FALCONEye's superior performance than the state-of-the-art in FALCON-Bench, and similar or better performance in related benchmarks.
This paper presents SANDMAN, an architecture for cyber deception that leverages Language Agents to emulate convincing human simulacra. Our 'Deceptive Agents' serve as advanced cyber decoys, designed for high-fidelity engagement with attackers by extending the observation period of attack behaviours. Through experimentation, measurement, and analysis, we demonstrate how a prompt schema based on the five-factor model of personality systematically induces distinct 'personalities' in Large Language Models. Our results highlight the feasibility of persona-driven Language Agents for generating diverse, realistic behaviours, ultimately improving cyber deception strategies.
Open-ended tasks are particularly challenging for LLMs due to the vast solution space, demanding both expansive exploration and adaptable strategies, especially when success lacks a clear, objective definition. Writing, with its vast solution space and subjective evaluation criteria, provides a compelling testbed for studying such problems. In this paper, we investigate the potential of LLMs to act as collaborative co-writers, capable of suggesting and implementing text improvements autonomously. We analyse three prominent LLMs - Gemini 1.5 Pro, Claude 3.5 Sonnet, and GPT-4o - focusing on how their action diversity, human alignment, and iterative improvement capabilities impact overall performance. This work establishes a framework for benchmarking autonomous writing agents and, more broadly, highlights fundamental challenges and potential solutions for building systems capable of excelling in diverse open-ended domains.