MOSAIC is a novel scheduling framework that significantly accelerates Mixture-of-Agents (MoA) workloads by jointly optimizing expert placement and utilizing confidence-aware adaptive aggregation.

RACE-Sched is an asynchronous agentic framework that successfully integrates low-latency, real-time scheduling decisions with advanced, long-horizon reasoning provided by Large Language Models.

CHRONOS is a novel three-layer architecture designed to address coupled failures in temporal data marketplaces by integrating temporal decay, changepoint-aware pricing, and differential privacy for ro…

The paper proposes a two-timescale governance framework using a multi-agent LLM to safely update and guide RL agents for SDN-IoT defense, significantly improving performance and stability under advers…

The paper proposes a policy-neutral execution and measurement layer to mediate between reinforcement learning policies and industrial environments, transforming ambiguous execution failures into struc…

The paper introduces AsyncTool, a new benchmark designed to evaluate LLM agents' ability to handle multiple, concurrent tasks with delayed tool feedback, demonstrating that asynchronous coordination i…

The paper proposes RoRo, a rubric-guided process reward framework that improves stepwise model routing by evaluating the quality of intermediate reasoning steps, leading to better performance and cost…

Lodestar is a novel online learning-based request routing system that significantly improves LLM inference efficiency by dynamically assigning incoming requests to the optimal GPU instance to minimize…

The paper introduces a queueing-theoretic framework to model dynamic cyber-attack surfaces, developing an adaptive reinforcement learning defense policy that significantly reduces active vulnerabiliti…

The paper demonstrates that for edge-native SLMs used in decentralized governance, simpler, intuitive reasoning (System 1) is significantly more robust and efficient than complex, iterative deliberati…

This paper introduces DIRECT, a routing framework that allocates test-time compute per prompt to improve the success--cost Pareto frontier for embodied agents.

The paper introduces FlowSteer, a prompt-only attack that exploits vulnerabilities in how multi-agent LLM systems plan workflows, significantly increasing the success rate of malicious signal propagat…

The paper proposes DecomposeR, a planner-centric framework that structures deep research into typed Directed Acyclic Graphs (DAGs) to explicitly improve the planning and execution of large language mo…

The paper proposes BaSE, a multi-armed bandit approach, to optimally allocate a fixed budget of LLM calls across parallel evolutionary search trajectories, significantly improving mean fitness and rel…

The paper proposes a system that operationalizes cybersecurity governance frameworks by integrating them with attack-path modeling and Deep Reinforcement Learning to generate practical, resource-const…

PlanTwin introduces a privacy-preserving architecture that allows cloud-hosted LLMs to plan over sensitive local environments by projecting the raw state into a sanitized, abstract digital twin.

DeepStage is a deep reinforcement learning framework that achieves autonomous, stage-aware defense against multi-stage APT campaigns by fusing graph-based telemetry and predicting attacker stages.

The paper argues that current 'on-the-fly' AI agent design lacks necessary software engineering rigor and proposes an 'AI Workflow Store' to provide hardened, reusable, and reliable agent workflows.

The paper proposes scheduling LLM agent workloads at the conversation level rather than the turn level, significantly reducing latency and improving energy efficiency by transforming unpredictable mul…

The Cognitive Firewall is a hybrid edge-cloud defense architecture that significantly reduces the attack success rate of Indirect Prompt Injection against browser-based AI agents by combining local vi…