This study investigates the implementation and efficacy of DeputyDev, an AI-powered code review assistant developed to address inefficiencies in the software development process. The process of code review is highly inefficient for several reasons, such as it being a time-consuming process, inconsistent feedback, and review quality not being at par most of the time. Using our telemetry data, we observed that at TATA 1mg, pull request (PR) processing exhibits significant inefficiencies, with average pick-up and review times of 73 and 82 hours, respectively, resulting in a 6.2 day closure cycle. The review cycle was marked by prolonged iterative communication between the reviewing and submitting parties. Research from the University of California, Irvine indicates that interruptions can lead to an average of 23 minutes of lost focus, critically affecting code quality and timely delivery. To address these challenges, we developed DeputyDev's PR review capabilities by providing automated, contextual code reviews. We conducted a rigorous double-controlled A/B experiment involving over 200 engineers to evaluate DeputyDev's impact on review times. The results demonstrated a statistically significant reduction in both average per PR (23.09%) and average per-line-of-code (40.13%) review durations. After implementing safeguards to exclude outliers, DeputyDev has been effectively rolled out across the entire organisation. Additionally, it has been made available to external companies as a Software-as-a-Service (SaaS) solution, currently supporting the daily work of numerous engineering professionals. This study explores the implementation and effectiveness of AI-assisted code reviews in improving development workflow timelines and code.

Multi-agent self-organizing systems (MASOS) exhibit key characteristics including scalability, adaptability, flexibility, and robustness, which have contributed to their extensive application across various fields. However, the self-organizing nature of MASOS also introduces elements of unpredictability in their emergent behaviors. This paper focuses on the emergence of dependency hierarchies during task execution, aiming to understand how such hierarchies arise from agents' collective pursuit of the joint objective, how they evolve dynamically, and what factors govern their development. To investigate this phenomenon, multi-agent reinforcement learning (MARL) is employed to train MASOS for a collaborative box-pushing task. By calculating the gradients of each agent's actions in relation to the states of other agents, the inter-agent dependencies are quantified, and the emergence of hierarchies is analyzed through the aggregation of these dependencies. Our results demonstrate that hierarchies emerge dynamically as agents work towards a joint objective, with these hierarchies evolving in response to changing task requirements. Notably, these dependency hierarchies emerge organically in response to the shared objective, rather than being a consequence of pre-configured rules or parameters that can be fine-tuned to achieve specific results. Furthermore, the emergence of hierarchies is influenced by the task environment and network initialization conditions. Additionally, hierarchies in MASOS emerge from the dynamic interplay between agents' "Talent" and "Effort" within the "Environment." "Talent" determines an agent's initial influence on collective decision-making, while continuous "Effort" within the "Environment" enables agents to shift their roles and positions within the system.

Vision Language Model (VLM)-based agents are stateful, autonomous entities capable of perceiving and interacting with their environments through vision and language. Multi-agent systems comprise specialized agents who collaborate to solve a (complex) task. A core security property is robustness, stating that the system should maintain its integrity under adversarial attacks. However, the design of existing multi-agent systems lacks the robustness consideration, as a successful exploit against one agent can spread and infect other agents to undermine the entire system's assurance. To address this, we propose a new defense approach, Cowpox, to provably enhance the robustness of multi-agent systems. It incorporates a distributed mechanism, which improves the recovery rate of agents by limiting the expected number of infections to other agents. The core idea is to generate and distribute a special cure sample that immunizes an agent against the attack before exposure and helps recover the already infected agents. We demonstrate the effectiveness of Cowpox empirically and provide theoretical robustness guarantees.

Future 6G radio access networks (RANs) will be artificial intelligence (AI)-native: observed, reasoned about, and re-configured by autonomous agents cooperating across the cloud-edge continuum. We introduce MX-AI, the first end-to-end agentic system that (i) instruments a live 5G Open RAN testbed based on OpenAirInterface (OAI) and FlexRIC, (ii) deploys a graph of Large-Language-Model (LLM)-powered agents inside the Service Management and Orchestration (SMO) layer, and (iii) exposes both observability and control functions for 6G RAN resources through natural-language intents. On 50 realistic operational queries, MX-AI attains a mean answer quality of 4.1/5.0 and 100 % decision-action accuracy, while incurring only 8.8 seconds end-to-end latency when backed by GPT-4.1. Thus, it matches human-expert performance, validating its practicality in real settings. We publicly release the agent graph, prompts, and evaluation harness to accelerate open research on AI-native RANs. A live demo is presented here: https://www.youtube.com/watch?v=CEIya7988Ug&t=285s&ab_channel=BubbleRAN

Current AI agents create significant barriers for users by requiring extensive processing to understand web pages, making AI-assisted web interaction slow and expensive. This paper introduces webMCP (Web Machine Context & Procedure), a client-side standard that embeds structured interaction metadata directly into web pages, enabling more efficient human-AI collaboration on existing websites. webMCP transforms how AI agents understand web interfaces by providing explicit mappings between page elements and user actions. Instead of processing entire HTML documents, agents can access pre-structured interaction data, dramatically reducing computational overhead while maintaining task accuracy. A comprehensive evaluation across 1,890 real API calls spanning online shopping, authentication, and content management scenarios demonstrates webMCP reduces processing requirements by 67.6% while maintaining 97.9% task success rates compared to 98.8% for traditional approaches. Users experience significantly lower costs (34-63% reduction) and faster response times across diverse web interactions. Statistical analysis confirms these improvements are highly significant across multiple AI models. An independent WordPress deployment study validates practical applicability, showing consistent improvements across real-world content management workflows. webMCP requires no server-side modifications, making it deployable across millions of existing websites without technical barriers. These results establish webMCP as a viable solution for making AI web assistance more accessible and sustainable, addressing the critical gap between user interaction needs and AI computational requirements in production environments.

Collective behaviors such as swarming and flocking emerge from simple, decentralized interactions in biological systems. Existing models, such as Vicsek and Cucker-Smale, lack collision avoidance, whereas the Olfati-Saber model imposes rigid formations, limiting their applicability in swarm robotics. To address these limitations, this paper proposes a minimal yet expressive model that governs agent dynamics using relative positions, velocities, and local density, modulated by two tunable parameters: the spatial offset and kinetic offset. The model achieves spatially flexible, collision-free behaviors that reflect naturalistic group dynamics. Furthermore, we extend the framework to cognitive autonomous systems, enabling energy-aware phase transitions between swarming and flocking through adaptive control parameter tuning. This cognitively inspired approach offers a robust foundation for real-world applications in multi-robot systems, particularly autonomous aerial swarms.

MicroRNAs (miRs) are robust regulators of gene expression, implicated in most biological processes. microRNAs predominantly downregulate the expression of genes post-transcriptionally and each miR is predicted to target several hundred genes. The accurate identification and annotation of miR-mRNA target interactions is central to understanding miRs function and their therapeutic potential. However, computational target prediction is challenging due to imperfect complementarity of miRs with their targets and the growing volume and heterogeneity of experimental data present challenges in accessing, integrating, and analysing miR-target interaction information across biological contexts. This creates a need for integrated resources and intelligent query tools. We present the miRKat Suite, comprising miRKatDB, a comprehensive, curated database of predicted and validated miR-target interactions and associated annotations, and miRKatAI, a multi-agent system powered by large language models (LLMs) and LangGraph. miRKatDB integrates data from multiple publicly available sources, providing a comprehensive foundation for miR studies, including miR target genes and changes in levels of tissue expression previously reported. miRKatAI offers a natural language interface for complex querying of miRKatDB, facilitates grounded information retrieval from established sources in the field, and supports basic data visualisation. The miRKat Suite aims to accelerate miR research by streamlining data access, enhancing exploratory analysis, and supporting hypothesis generation.

Hegselmann--Krause models are localized, distributed averaging dynamics on spatial data. A key aspect of these dynamics is that they lead to cluster formation, which has important applications in geographic information systems, dynamic clustering algorithms, opinion dynamics, and social networks. For these models, the key questions are whether a fixed point exists and, if so, characterizing it. In this work, we establish new results towards the "2R-Conjecture" for the Hegselmann--Krause model, for which no meaningful progress, or even any precise statement, has been made since its introduction in 2007. This conjecture relates to the structure of the fixed point when there are a large number of agents per unit space. We provide, among other results, a proof in expectation and a statement of a stronger result that is supported by simulation. The key methodological contribution is to consider the dynamics as an infinite-dimensional problem on the space of point processes, rather than on finitely many points. This enables us to leverage stationarity, shift invariance, and certain other symmetries to obtain the results. These techniques do not have finite-dimensional analogs.

Recent advancements in embodied AI and robotic manipulation have highlighted the need for scalable, interactive simulation environments that support both scene construction and data collection for training autonomous agents. While existing platforms [1, 2, 3, 4] offer partial solutions, such as scene generation[5, 3, 4] or task-specific manipulation datasets [6, 7, 8], few provide a unified framework that integrates high-fidelity scene construction with flexible mobile robotic data collection system. To bridge this gap, we present AgentWorld, an interactive simulation platform designed for procedural scene construction and mobile-based teleoperation, enabling efficient data collection for imitation learning in complex household environments. AgentWorld addresses two critical challenges in embodied AI research: (1) stable and diverse scene generation, ensuring that the simulated environments are visually realistic and physically plausible, and (2) a comprehensive data collection system in simulation, which allows seamless control of mobile bases and robotic arms for data collection. AgentWorld is built upon NVIDIA's Omniverse Isaac Sim [9] and Unreal Engine [10], allowing it to inherit both strengths including the physics engine for robot parallel training and realistic rendering effects.

Recent progress in large language models (LLMs) has enabled the development of autonomous web agents capable of navigating and interacting with real websites. However, evaluating such agents remains challenging due to the instability and inconsistency of existing benchmarks, which often rely on dynamic content or oversimplified simulations. In this work, we introduce WebArXiv, a static and time-invariant benchmark comprising 275 web-based tasks grounded in the arXiv platform. WebArXiv ensures reproducible and reliable evaluation by anchoring tasks in fixed web snapshots with deterministic ground truths and standardized action trajectories. Through behavioral analysis, we identify a common failure mode, Rigid History Reflection, where agents over-rely on fixed interaction histories. To address this, we propose a lightweight dynamic reflection mechanism that allows agents to selectively retrieve relevant past steps during decision-making. We evaluate ten state-of-the-art web agents on WebArXiv. Results demonstrate clear performance differences across agents and validate the effectiveness of our proposed reflection strategy.