This paper investigates the effectiveness of small language models (SLMs) for agentic tasks (function/tool/API calling) with a focus on running agents on edge devices without reliance on cloud infrastructure. We evaluate SLMs using the Berkeley Function Calling Leaderboard (BFCL) framework and describe parameter-driven optimization strategies that include supervised fine-tuning (SFT), parameter-efficient fine-tuning (PEFT), reinforcement learning (RL)-based optimization, preference alignment via Direct Preference Optimization (DPO), and hybrid methods. We report results for models including TinyAgent, TinyLlama, Qwen, and xLAM across BFCL categories (simple, multiple, parallel, parallel-multiple, and relevance detection), both in live and non-live settings, and in multi-turn evaluations. We additionally detail a DPO training pipeline constructed from AgentBank data (e.g., ALFRED), including our conversion of SFT data to chosen-rejected pairs using TinyLlama responses as rejected outputs and manual validation. Our results demonstrate clear accuracy differences across model scales where medium-sized models (1-3B parameters) significantly outperform ultra-compact models (<1B parameters), achieving up to 65.74% overall accuracy, and 55.62% multi-turn accuracy with hybrid optimization. This study highlights the importance of hybrid optimization strategies that enable small language models to deliver accurate, efficient, and stable agentic AI on edge devices, making privacy-preserving, low-latency autonomous agents practical beyond the cloud.

Dynamic sleep mode optimization (SMO) in millimeter-wave (mmWave) networks is essential for maximizing energy efficiency (EE) under stringent quality-of-service (QoS) constraints. However, existing optimization and reinforcement learning (RL)-based approaches rely on aggregated, static base station (BS) traffic models that fail to capture non-stationary traffic dynamics and suffer from prohibitively large state-action spaces, limiting their real-world deployment. To address these challenges, this paper proposes a Multi-Agent Deep Reinforcement Learning (MARL) framework employing a Double Deep Q-Network (DDQN), referred to as MARL-DDQN, for adaptive SMO in a 3D urban environment using a time-varying and community-based user equipment (UE) mobility model. Unlike conventional single-agent RL, the proposed MARL-DDQN enables scalable, distributed decision-making with minimal signaling overhead. A realistic BS power consumption model and beamforming are integrated to accurately quantify EE, while QoS is uniquely defined in terms of throughput. The proposed method adaptively learns SMO policies to maximize EE while mitigating inter-cell interference and ensuring throughput fairness. Extensive simulations demonstrate that MARL-DDQN consistently outperforms state-of-the-art SM strategies, including the All On, iterative QoS-aware load-based (IT-QoS-LB), MARL-DDPG, and MARL-PPO, achieving up to 0. 60 Mbit/Joule EE, 8. 5 Mbps 10th-percentile throughput, and satisfying QoS constraints 95 % of the time under dynamic network scenarios. I. Introduction The exponential growth in mobile data demand has necessitated increased spectrum availability and accelerated the expansion of cellular network infrastructure. To address the limitations of the sub-6 GHz spectrum, millimeter wave (mmWave) communications, operating within the 30-300 GHz band, have emerged as a key enabler in fifth-generation (5G) networks. With significantly larger bandwidth availability, mmWave technology presents a viable solution to spectrum scarcity challenges [1]. However, mmWave signals suffer from high propagation loss, atmospheric absorption, and susceptibility to blockages, which severely limit coverage and reliability. To address coverage and growing capacity demands, 5G networks rely on densification, deploying numerous low-power mmWave BSs with inter-site distances of a few hundred meters [1]. These BSs utilize large antenna arrays to enable beamforming and spatial multiplexing, often relying on hybrid analog-digital precoding to reduce hardware complexity [2]. However, the RF chain remains a major source of power consumption, particularly the Analog-to-digital converters (ADCs) and digital-to-analog converters (DACs), whose power scales with sampling rate. Due to the higher frequencies and wider bandwidths of mmWave systems, these components require significantly higher sampling rates than sub-6 GHz systems [3], resulting in substantial energy demands.

The Internet of Agents (IoA) is rapidly gaining prominence as a foundational architecture for interconnected intelligent systems, designed to facilitate seamless discovery, communication, and collaborative reasoning among a vast network of Artificial Intelligence (AI) agents. Powered by Large Language and Vision-Language Models, IoA enables the development of interactive, rational agents capable of complex cooperation, moving far beyond traditional isolated models. IoA involves physical entities, i.e., Wireless Agents (WAs) with limited onboard resources, which need to offload their compute-intensive agentic AI services to nearby servers. Such servers can be Mobile Agents (MAs), e.g., vehicle agents, or Fixed Agents (FAs), e.g., end-side units agents. Given their fixed geographical locations and stable connectivity, FAs can serve as reliable communication gateways and task aggregation points. This stability allows them to effectively coordinate with and offload to an Aerial Agent (AA) tier, which has an advantage not affordable for highly mobile MAs with dynamic connectivity limitations. As such, we propose a two-tier optimization approach. The first tier employs a multi-leader multi-follower Stackelberg game. In the game, MAs and FAs act as the leaders who set resource prices. WAs are the followers to determine task offloading ratios. However, when FAs become overloaded, they can further offload tasks to available aerial resources. Therefore, the second tier introduces a Double Dutch Auction model where overloaded FAs act as the buyers to request resources, and AAs serve as the sellers for resource provision. We then develop a diffusion-based Deep Reinforcement Learning algorithm to solve the model. Numerical results demonstrate the superiority of our proposed scheme in facilitating task offloading.

This paper introduces a dynamic and actionable framework for securing agentic AI systems in enterprise deployment. We contend that safety and security are not merely fixed attributes of individual models but also emergent properties arising from the dynamic interactions among models, orchestrators, tools, and data within their operating environments. We propose a new way of identification of novel agentic risks through the lens of user safety. Although, for traditional LLMs and agentic models in isolation, safety and security has a clear separation, through the lens of safety in agentic systems, they appear to be connected. Building on this foundation, we define an operational agentic risk taxonomy that unifies traditional safety and security concerns with novel, uniquely agentic risks, including tool misuse, cascading action chains, and unintended control amplification among others. At the core of our approach is a dynamic agentic safety and security framework that operationalizes contextual agentic risk management by using auxiliary AI models and agents, with human oversight, to assist in contextual risk discovery, evaluation, and mitigation. We further address one of the most challenging aspects of safety and security of agentic systems: risk discovery through sandboxed, AI-driven red teaming. We demonstrate the framework effectiveness through a detailed case study of NVIDIA flagship agentic research assistant, AI-Q Research Assistant, showcasing practical, end-to-end safety and security evaluations in complex, enterprise-grade agentic workflows. This risk discovery phase finds novel agentic risks that are then contextually mitigated. We also release the dataset from our case study, containing traces of over 10,000 realistic attack and defense executions of the agentic workflow to help advance research in agentic safety.

Feature transformation enhances downstream task performance by generating informative features through mathematical feature crossing. Despite the advancements in deep learning, feature transformation remains essential for structured data, where deep models often struggle to capture complex feature interactions. Prior literature on automated feature transformation has achieved success but often relies on heuristics or exhaustive searches, leading to inefficient and time-consuming processes. Recent works employ reinforcement learning (RL) to enhance traditional approaches through a more effective trial-and-error way. However, two limitations remain: 1) Dynamic feature expansion during the transformation process, which causes instability and increases the learning complexity for RL agents; 2) Insufficient cooperation and communication between agents, which results in suboptimal feature crossing operations and degraded model performance. To address them, we propose a novel heterogeneous multi-agent RL framework to enable cooperative and scalable feature transformation. The framework comprises three heterogeneous agents, grouped into two types, each designed to select essential features and operations for feature crossing. To enhance communication among these agents, we implement a shared critic mechanism that facilitates information exchange during feature transformation. To handle the dynamically expanding feature space, we tailor multi-head attention-based feature agents to select suitable features for feature crossing. Additionally, we introduce a state encoding technique during the optimization process to stabilize and enhance the learning dynamics of the RL agents, resulting in more robust and reliable transformation policies. Finally, we conduct extensive experiments to validate the effectiveness, efficiency, robustness, and interpretability of our model.

Emotion-aware customer service needs in-domain conversational data, rich annotations, and predictive capabilities, but existing resources for emotion recognition are often out-of-domain, narrowly labeled, and focused on post-hoc detection. To address this, we conducted a controlled Wizard of Oz (WOZ) experiment to elicit interactions with targeted affective trajectories. The resulting corpus, EmoWOZ-CS, contains 2,148 bilingual (Dutch-English) written dialogues from 179 participants across commercial aviation, e-commerce, online travel agencies, and telecommunication scenarios. Our contributions are threefold: (1) Evaluate WOZ-based operator-steered valence trajectories as a design for emotion research; (2) Quantify human annotation performance and variation, including divergences between self-reports and third-party judgments; (3) Benchmark detection and forward-looking emotion inference in real-time support. Findings show neutral dominates participant messages; desire and gratitude are the most frequent non-neutral emotions. Agreement is moderate for multilabel emotions and valence, lower for arousal and dominance; self-reports diverge notably from third-party labels, aligning most for neutral, gratitude, and anger. Objective strategies often elicit neutrality or gratitude, while suboptimal strategies increase anger, annoyance, disappointment, desire, and confusion. Some affective strategies (cheerfulness, gratitude) foster positive reciprocity, whereas others (apology, empathy) can also leave desire, anger, or annoyance. Temporal analysis confirms successful conversation-level steering toward prescribed trajectories, most distinctly for negative targets; positive and neutral targets yield similar final valence distributions. Benchmarks highlight the difficulty of forward-looking emotion inference from prior turns, underscoring the complexity of proactive emotion-aware support.

Abstract--The Multi-Agent Path Finding (MAPF) problem aims to find collision-free paths for multiple agents while optimizing objectives such as the sum of costs or makespan. MAPF has wide applications in domains like automated warehouses, manufacturing systems, and airport logistics. However, most MAPF formulations assume a simplified robot model for planning, which overlooks execution-time factors such as kinodynamic constraints, communication latency, and controller variability. This gap between planning and execution is problematic for time-sensitive applications. T o bridge this gap, we propose REMAP, an execution-informed MAPF planning framework that can be combined with leading search-based MAPF planners with minor changes. Our framework integrates the proposed ExecTimeNet to accurately estimate execution time based on planned paths. We demonstrate our method for solving MAPF with Real-world Deadlines (MAPF-RD) problem, where agents must reach their goals before a predefined wall-clock time. We integrate our framework with two popular MAPF methods, MAPF-LNS and CBS. Experiments show that REMAP achieves up to 20% improvement in solution quality over baseline methods (e.g., constant execution speed estimators) on benchmark maps with up to 300 agents. The Multi-Agent Path Finding (MAPF) problem seeks to find collision-free paths for multiple agents in a shared environment while optimizing objectives such as the sum of costs or makespan. This problem is a fundamental challenge in settings such as automated warehouses [1], traffic intersections [2], and airport logistics [3]. State-of-the-art MAPF methods can efficiently coordinate hundreds of agents, making MAPF a promising solution for these domains.

We study whether large language models acting as autonomous bidders can tacitly collude by coordinating when to accept platform posted payouts in repeated Dutch auctions, without any communication. We present a minimal repeated auction model that yields a simple incentive compatibility condition and a closed form threshold for sustainable collusion for subgame-perfect Nash equilibria. In controlled simulations with multiple language models, we observe systematic supra-competitive prices in small auction settings and a return to competitive behavior as the number of bidders in the market increases, consistent with the theoretical model. We also find LLMs use various mechanisms to facilitate tacit coordination, such as focal point acceptance timing versus patient strategies that track the theoretical incentives. The results provide, to our knowledge, the first evidence of bidder side tacit collusion by LLMs and show that market structure levers can be more effective than capability limits for mitigation.

The risk of indoor airborne transmission among co-located individuals is generally non-uniform, which remains a critical challenge for public health modelling. Thus, we present CompARE, an integrated risk assessment framework for indoor airborne disease transmission that reveals a striking bimodal distribution of infection risk driven by airflow dynamics and human behavior. Combining computational fluid dynamics (CFD), machine learning (ML), and agent-based modeling (ABM), our model captures the complex interplay between aerosol transport, human mobility, and environmental context. Based on a prototypical childcare center, our approach quantifies how incorporation of ABM can unveil significantly different infection risk profiles across agents, with more than two-fold change in risk of infection between the individuals with the lowest and highest risks in more than 90% of cases, despite all individuals being in the same overall environment. We found that infection risk distributions can exhibit not only a striking bimodal pattern in certain activities but also exponential decay and fat-tailed behavior in others. Specifically, we identify low-risk modes arising from source containment, as well as high-risk tails from prolonged close contact. Our approach enables near-real-time scenario analysis and provides policy-relevant quantitative insights into how ventilation design, spatial layout, and social distancing policies can mitigate transmission risk. These findings challenge simple distance-based heuristics and support the design of targeted, evidence-based interventions in high-occupancy indoor settings.

Current AI agents excel in familiar settings, but fail sharply when faced with novel tasks with unseen vocabularies -- a core limitation of procedural memory systems. We present the first benchmark that isolates procedural memory retrieval from task execution, evaluating whether agents can recognize functionally equivalent procedures that span different object instantiations. Using ALFWorld, we construct dual corpora of expert and LLM-generated trajectories and evaluate six retrieval methods using systematically stratified queries. Our results expose a clear generalization cliff: embedding-based methods perform strongly on familiar contexts, yet degrade considerably on novel ones, while LLM-generated procedural abstractions demonstrate reliable cross-context transfer. Controlled ablations show that although embeddings capture some lexical-level abstraction, they fundamentally treat procedures as unordered bags of words, discarding temporal structure necessary for cross-context transfer. Corpus scale delivers far larger gains than representation enrichment, revealing an architectural ceiling in current encoders. Our benchmark offers the first diagnostic framework separating genuine procedural understanding from surface-level memorization and gives tools for developing retrieval systems capable of dependable generalization. Resources available at our GitHub repository (https://github.com/qpiai/Proced_mem_bench).