...SWE-agent: ...Agent-Computer Interfaces Enable Automated Software Engineering John Yang Carlos E. Jimenez Alexan...affects the performance of language model agents. As a result of this exploration, we introduce SWE-agent: a system that facilitates LM agents to autonomously use computers to solve software engineering ta...

...Consider the simple setting of an agent interacting directly with a Linux shell [59]. In practice, we find that LM agents can struggle to r... makes an invalid edit. These deficits substantially hamper performance, motivating the need for an agent-computer interface (ACI), i.e., an abstraction layer between the LM ...agent and computer, to enhance the LM ...

...ce that was built informed by these differences. Ultimately, a well-designed ACI should help the LM agent understand the state of the application given previous changes, manage history to avoid unnecessary...rmance [7]. We use two approaches to enhance performance on a development set: (1) manually inspect agent behavior to identify difficulties and propose improvements, and (2) run a grid search to select the... use, reducing the need for demonstrations or fine-tuning. This is a defining principle for all SWE-agent commands that we describe in Section 3. 2. Actions should be compact and efficient. Important opera...

...DialogAgent: An Auto-engagement ...Agent for Code Question Answering Data Production Xiaoyun Liang Jingyi Ren Jiayi Qi ByteDance ByteDan...specific user and business contexts, accessible, labeled datasets. In this paper, we present DialogAgent, an automated tool for generating synthetic training data all while adhering to strict privacy pol...

... GENT datasets and scale effectively to handle complex tasks. In this section, we introduce DialogAgent, a novel method Encoder-only models, such as CodeBERT [11], rely on designed to automatically ge... to learn vectorized embeddings of input code IDE usage. Figure 1 illustrates the workflow of DialogAgent, sequences. These models are particularly suited for tasks which leverages the Q&A plugin within th...odex [14], utilize an autoregressive transformer reflects the day-to-day needs of developers. DialogAgent generates decoder to generate code sequences. These models excel synthetic data by simulating multi...

...Fig. 1: Workflow of DialogAgent Fig. 2: Process of QA-DBA Fig. 4: Prompt for Data Production Planning. B. Chat Configuration Gen...

...Abstract In this article, I describe agent-centered search (also called real-time search or local search) and illustrate this planning paradig...m with examples. Agent-centered search methods interleave planning and plan execution and restrict planning to the part of... the domain around the current state of the agent, for example, the current location of a mobile robot or the current board position of a game. These...

...AI Magazine Volume 22 Number 4 (2001) ( AAAI) Articles Agent-Centered Search Sven Koenig In this article, I describe ...agent-centered search (also called real-time search or local search) and illustrate this planning paradig...m with examples. Agent-centered search methods interleave planning and plan execution and restrict planning to the part of...

...Articles local search space current state Figure 1. Agent-Centered Search. consumption) and then execute them. Thus, they are offline planning methods. ...Agent-centered search methods, however, interleave planning and execution and are, thus, online planning ...e constraints and (2) they often decrease the sum of planning and execution cost. Time constraints: Agent-centered search methods can execute actions in the presence of soft or hard time constraints. The p...

...Articles traditional search planning plan execution agent-centered search small (bounded) planning cost between plan executions small sum of planning and exe...cution cost Figure 2. Traditional Search versus Agent-Centered Search. cially because determining plans with minimal execution cost is often intractable,...an can be determined automatically (even dynamically), using either techniques tailored to specific agent-centered search methods (Ishida 1992) or general techniques from limited rationality and deliberati...

...AI Magazine Volume 24 Number 1 (2003) ( AAAI) Articles The 2002 Trading Agent Competition An Overview of ...Agent Strategies Amy Greenwald This article summarizes 16 ...agent strategies that were designed for the 2002 Trading ...

...Articles Agent AI Technology Institution ATT AC Machine Learning ATT Labs, Research AVAYAH Neural Networks Oracl...e LIVINGAGENTS Multiagent Systems Living Systems, AG PACKATAC A Decision-Theoretic Agent North Carolina State University ROXYBOT Online Heuristic Search Brown University PAININNEC Geneti...c Algorithms NEC Research OUTHAMPTONTAC An Adaptive Agent University of Southampton THALIS Risk Aversion University of Essex TOMAHACK POMDPs University of ...

...y, SICS also developed new TAC A GENT-WARE in the form of a JAVA software tool kit for GENTWARE. Agent Strategies This section describes 16 TAC ...agent strategies. 2 Fifteen of these 16 agents participated in the semifinals. A list of the 16 agent...le 1, along with the primary AI technology that characterizes each one. ATTAC: A Machine Learning Agent The core of ATTAC'S approach is a learning algorithm ATTAC-02 is essentially the ATTAC-01 (Stone...

...Explainability in Human-Agent Systems A Preprint Avi Rosenfeld Department of Computer Science of the Jerusalem College of Techn...ards@jct.ac.il April 18, 2019 Abstract This paper presents a taxonomy of explainability in Human-Agent Systems. We consider fundamental questions about the Why, Who, What, When and How of explainability...l need to evaluate all other issues regarding explainability. Keywords XAI · Explainability · Human-Agent Systems 1 Introduction As the field of Artificial Intelligence matures and becomes ubiquitous, th...

...If the system is agent-centric, such as in knowledge discovery or self-driving cars, the ...agent might need to provide information about its decision to help convince the human participant of the ...correctness of their solution, aiding in the adoption of these agent based technologies [102]. In both cases, the information the ...

...ded to address the receiver's request. 2.2 The Need for Precisely Defining Explainability in Human-Agent Systems Recently, questions have arose as to the definition of explainability of machine learning a...nd agent systems. An explosive growth of interest has been registered within various research communities as...ainable systems, as both terms refer to aspects of providing information to a human actor about the agent's decision-making process. Previous work generally defined interpretability as the ability to expla...

...The AI Agent Index Stephen Casper * 1 Luke Bailey 2 Rosco Hunter 3 Carson Ezell 4 Emma Cabalé 5 Micha...ems. To fill this gap, we introduce broader swathes of economic activity (Wang et al., 2024; the AI Agent Index, the first public database to Durante et al., 2024; Sager et al., 2025). As a result, their d... safety and visibility into their operation (Shavit et al., 2023; risk management practices. The AI Agent Index Chan et al., 2024a;b; 2025; Kolt, 2025), many practical is available online at https://aiagen...

...The AI Agent Index Publicly available formal 19.4% safety policy Publicly reported external 7.5% safety testing...er 0 20 40 60 80 100 0 25 50 75 100 % Publicly Available % With Safety Measure Figure 1. Most AI agent developers in the index provide some Figure 2. Only 19.4% of indexed agentic systems disclose a for...d country of origin). 1989; Dung, 2024). While there have been notable attempts to define the term "agent", including in the context In addition to collecting and systematizing information of computational...

...The AI Agent Index the base model and/or in external storage modules (Sumers datasheets (Gebru et al., 2018), m...3; Wu et al., 2024). Progress in these applications is being evaluated by a growing suite of The AI Agent Index fills this gap. By collecting and communicating benchmarks, which measure performance in comp...g websites) (Fang et al., 2024a; Jaech et al., 2024). there is no widely-accepted definition of "AI agent." We Additionally, as agentic AI systems undertake more complex do not propose one here. Given our ...

...Context is Key for Agent Security Lillian Tsai Eugene Bagdasarian Google, Inc. Google Research tslilyai@google.com ebag...contexts, system like a filesystem or OS, have become increasingly particularly for general-purpose agent systems; important to address with the rise of systems like generalist (AI) agents [3, 23, 55, 69, ...

...Lillian Tsai and Eugene Bagdasarian (2) Conseca, a contextually-aware agent security framework User Task, that scales security for generalist agents using Trusted + Untrust...enforced policies specific Executes Potentially to a particular purpose and trusted context; and Agent Harmful Actions (3) A proof-of-concept prototype of Conseca integrated Tools with a Linux compu...ter-use agent ( 4, 5). User's ...

...Context is Key for Agent Security content (e.g., email or file contents). The planner returns to (2) on action type or argu...ontext, or denying an action, Conseca returns the rationale for and actions/outputs history. If the agent has completed the the decision to the ...agent for transparency and feedback. user's task, the user receives the final output. 3.4 Security Discus...

...Private Agent-Based Modeling Ayush Chopra Arnau Quera-Bofarull Nurullah Giray-Kuru Massachusetts Institute of T...askar@media.mit.edu ABSTRACT is hindered by (1) the need for microdata to generate the underlying agent population and (2) the often substantial computational The practical utility of ...agent-based models in decision-making relies resources required to run, calibrate, and analyze an ABM. Re...

...y infection or mortality data. It is important to recognize that due to the stochasticity of the 2 AGENT-BASED MODELS model and its partial observability, multiple sets of parameter values In this sectio...We denote by met by adopting a Bayesian framework, wherein parameter inference z ( ) the state of agent at time, which encapsulates both fixed corresponds to determining the posterior distribution over t...the account balance of firms in a financial (y |) () auction model. As the simulation proceeds, an agent updates (| y) = , (2) (y) their state z ( ) by interacting with their neighbors N ( ) and the ...

...l the personal such as stochastic gradient descent to minimize L(). Given the health information of agent cohorts. In the context of ...agent-based differentiablity of the ABM, the gradient of the expected loss can be modeling, this informat...ational family Q will be parameterized by a adversarial participants. These participants, either an agent or the deep neural network (i.e, a normalizing flow) that is trained using server, aim to learn pri...

...An Agent-based Classification Model Feng Gu, Uwe Aickelin, Julie Greensmith School of Computer Science Uni... It has been successfully applied to intrusion detection in computer security. It is believed that agent-based modelling is an ideal approach for implementing AIS, as intelligent agents could be the perfe...ns of immune entities in AIS. This model evaluates the feasibility of re-implementing the DCA in an agent-based simulation environment called AnyLogic, where the immune entities in the DCA are represented ...

... 4. return mature context 1. acquire data items 3. acquire attributes Antigen agents DC agents TC agent 2. 'picked' message 6. generate MCAV diagram and true positive 5. acquire MCAV, original category...Three types of agents are created in this model which are Antigen agents, DC agents and TC (T-cell) agent, to represent the entities involved in the DCA. Antigen agents are the data carriers, DC agents are... the data processors and TC agent is the statistical analyser of the classifi cation. The TC ...