Deployment-Time Memorization in Foundation-Model Agents

Foundation-model agents are increasingly long-lived systems that remember users across interactions, making memorization an explicit deployment-time function rather than solely a property of model weights. Existing work addresses parametric memorization or audits fixed memory configurations, but does not characterize how memory-design choices jointly shape personalization utility, extraction risk, and deletion fidelity. We study this surface as deployment-time memorization, formulating agent memory as a privacy-utility frontier measured by Personalization Recall (PR) and Adversarial Extraction Rate (AER), and sweeping three memory-design knobs: summarization aggressiveness, retrieval breadth (k), and deletion mode. We further introduce the Forgetting Residue Score (FRS) to quantify whether deleted information remains recoverable from derived memory tiers. On LongMemEval, key-fact summarization reduces canary extraction by 76% on Gemma 3 12B and 64% on GPT-4o-mini while preserving nearly all personalization recall; critically, once content is compressed away, increasing k no longer restores leakage. The same compression, however, induces a deletion-fidelity failure: raw-only deletion leaves derived summary copies recoverable in approximately 20% of instances, and only full-pipeline purge or tombstone redaction drives worst-tier residue to zero. Together, these results establish that persistent agent memory must be evaluated as a first-class memorization mechanism -- assessed by what it helps agents recall, what it makes extractable, and what it can truly erase.

EvalStop: Using World Feedback to Detect and Correct Reward Overoptimization in Multi-Tenant RLHF Platforms

Cloud LLM fine-tuning platforms increasingly serve RLHF workloads, where a learned reward model is optimized as a proxy for human quality. As Gao et al. (2023) showed, this proxy diverges from world feedback (downstream eval metrics) under sustained optimization pressure, a phenomenon known as reward overoptimization. Existing platform schedulers ignore this divergence: non-clairvoyant schedulers optimize JCT without any quality signal, SLAQ-style quality-aware schedulers use training loss (a weaker proxy that drops monotonically through hacking), and classical per-job early stopping requires human monitoring and does not free shared GPUs. We propose EvalStop, a composable scheduling primitive that terminates jobs on k consecutive eval-score declines, releases GPUs, preserves the best checkpoint, and delegates to any base scheduler. We frame scheduler-level early stopping as a detection problem and evaluate it in a discrete-event simulator whose RLHF workload mixes reward-hacking and structurally healthy runs, with ground-truth labels hidden from schedulers. On RLHF-heavy workloads (80% RLHF, 64 GPUs), EvalStop achieves precision 98% / recall 99% / FPR 1.5% while improving JCT by 9% and cutting wasted compute by 22% over SRTF-Est (p<0.05). Trivial fixed-progress and loss-plateau competitors either incur 65% FPR on healthy RLHF or miss over half of true hacking cases. Gains compose across every base scheduler tested (9-25% JCT) and detection quality stays stable under eval noise (precision at least 91% at noise std <= 0.05) and hacking base rate (precision at least 89% across 20-80% hacking fractions).

When Does Deep RL Beat Calibrated Baselines? A Benchmark Study on Adaptive Resource Control

A properly calibrated rule-based autoscaler can beat every one of six mainstream deep reinforcement learning (DRL) algorithms on cost across every workload we test - so when, if ever, does DRL actually help? We study this in RLScale-Bench, a reproducible benchmark and evaluation protocol for DRL on adaptive resource control, where an agent allocates compute to a dynamic workload under cost and service-level constraints. We evaluate PPO, DQN, A2C, SAC, TD3, and DDPG under matched architectures, training budgets, and reward functions against a calibrated rule-based baseline across six workload patterns and five seeds (240 runs), instantiate the benchmark on Kubernetes Horizontal Pod Autoscaling, and probe distribution-shift generalization. Three findings challenge common assumptions: (i) the calibrated controller achieves the lowest cost on all six workloads, though it trails the best RL agents on bursty and flash traffic; (ii) discrete-action algorithms outperform continuous-action ones by one to two orders of magnitude in constraint violations due to action-space mismatch; and (iii) no single algorithm dominates across workloads, with rankings shifting by up to four positions. The bottleneck in RL-based resource control is not algorithm selection but baseline calibration, reward engineering, and realistic evaluation protocols.

Robust Explanations for User Trust in Enterprise NLP Systems

Robust explanations are increasingly required for user trust in enterprise NLP, yet pre-deployment validation is difficult in the common case of black-box deployment (API-only access) where representation-based explainers are infeasible and existing studies provide limited guidance on whether explanations remain stable under real user noise, especially when organizations migrate from encoder classifiers to decoder LLMs. To close this gap, we propose a unified black-box robustness evaluation framework for token-level explanations based on leave-one-out occlusion, and operationalize explanation robustness with top-token flip rate under realistic perturbations (swap, deletion, shuffling, and back-translation) at multiple severity levels. Using this protocol, we conduct a systematic cross-architecture comparison across three benchmark datasets and six models spanning encoder and decoder families (BERT, RoBERTa, Qwen 7B/14B, Llama 8B/70B; 64,800 cases). We find that decoder LLMs produce substantially more stable explanations than encoder baselines (73% lower flip rates on average), and that stability improves with model scale (44% gain from 7B to 70B). Finally, we relate robustness improvements to inference cost, yielding a practical cost-robustness tradeoff curve that supports model and explanation selection prior to deployment in compliance-sensitive applications.

CarbonEdge: Carbon-Aware Deep Learning Inference Framework for Sustainable Edge Computing

Deep learning applications at the network edge lead to a significant growth in AI-related carbon emissions, presenting a critical sustainability challenge. The existing edge computing frameworks optimize for latency and throughput, but they largely ignore the environmental impact of inference workloads. This paper introduces CarbonEdge, a carbon-aware deep learning inference framework that extends adaptive model partitioning with carbon footprint estimation and green scheduling apabilities. We propose a carbon-aware scheduling algorithm that extends traditional weighted scoring with a carbon efficiency metric, supporting a tunable performance--carbon trade-off (demonstrated via weight sweep). Experimental evaluations on Docker-simulated heterogeneous edge environments show that CarbonEdge-Green mode achieves a 22.9% reduction in carbon emissions compared to monolithic execution. The framework achieves 1.3x improvement in carbon efficiency (245.8 vs 189.5 inferences per gram CO2) with negligible scheduling overhead (0.03ms per task). These results highlight the framework's potential for sustainable edge AI deployment, providing researchers and practitioners a tool to quantify and minimize the environmental footprint of distributed deep learning inference.

Adaptive Memory Admission Control for LLM Agents

LLM-based agents increasingly rely on long-term memory to support multi-session reasoning and interaction, yet current systems provide little control over what information is retained. In practice, agents either accumulate large volumes of conversational content, including hallucinated or obsolete facts, or depend on opaque, fully LLM-driven memory policies that are costly and difficult to audit. As a result, memory admission remains a poorly specified and weakly controlled component in agent architectures. To address this gap, we propose Adaptive Memory Admission Control (A-MAC), a framework that treats memory admission as a structured decision problem. A-MAC decomposes memory value into five complementary and interpretable factors: future utility, factual confidence, semantic novelty, temporal recency, and content type prior. The framework combines lightweight rule-based feature extraction with a single LLM-assisted utility assessment, and learns domain-adaptive admission policies through cross-validated optimization. This design enables transparent and efficient control over long-term memory. Experiments on the LoCoMo benchmark show that A-MAC achieves a superior precision-recall tradeoff, improving F1 to 0.583 while reducing latency by 31% compared to state-of-the-art LLM-native memory systems. Ablation results identify content type prior as the most influential factor for reliable memory admission. These findings demonstrate that explicit and interpretable admission control is a critical design principle for scalable and reliable memory in LLM-based agents.

LLMs as Orchestrators: Constraint-Compliant Multi-Agent Optimization for Recommendation Systems

Recommendation systems must optimize multiple objectives while satisfying hard business constraints such as fairness and coverage. For example, an e-commerce platform may require every recommendation list to include items from multiple sellers and at least one newly listed product; violating such constraints--even once--is unacceptable in production. Prior work on multi-objective recommendation and recent LLM-based recommender agents largely treat constraints as soft penalties or focus on item scoring and interaction, leading to frequent violations in real-world deployments. How to leverage LLMs for coordinating constrained optimization in recommendation systems remains underexplored. We propose DualAgent-Rec, an LLM-coordinated dual-agent framework for constrained multi-objective e-commerce recommendation. The framework separates optimization into an Exploitation Agent that prioritizes accuracy under hard constraints and an Exploration Agent that promotes diversity through unconstrained Pareto search. An LLM-based coordinator adaptively allocates resources between agents based on optimization progress and constraint satisfaction, while an adaptive epsilon-relaxation mechanism guarantees feasibility of final solutions. Experiments on the Amazon Reviews 2023 dataset demonstrate that DualAgent-Rec achieves 100% constraint satisfaction and improves Pareto hypervolume by 4-6% over strong baselines, while maintaining competitive accuracy-diversity trade-offs. These results indicate that LLMs can act as effective orchestration agents for deployable and constraint-compliant recommendation systems.

RobustExplain: Evaluating Robustness of LLM-Based Explanation Agents for Recommendation

Large Language Models (LLMs) are increasingly used to generate natural-language explanations in recommender systems, acting as explanation agents that reason over user behavior histories. While prior work has focused on explanation fluency and relevance under fixed inputs, the robustness of LLM-generated explanations to realistic user behavior noise remains largely unexplored. In real-world web platforms, interaction histories are inherently noisy due to accidental clicks, temporal inconsistencies, missing values, and evolving preferences, raising concerns about explanation stability and user trust. We present RobustExplain, the first systematic evaluation framework for measuring the robustness of LLM-generated recommendation explanations. RobustExplain introduces five realistic user behavior perturbations evaluated across multiple severity levels and a multi-dimensional robustness metric capturing semantic, keyword, structural, and length consistency. Our goal is to establish a principled, task-level evaluation framework and initial robustness baselines, rather than to provide a comprehensive leaderboard across all available LLMs. Experiments on four representative LLMs (7B--70B) show that current models exhibit only moderate robustness, with larger models achieving up to 8% higher stability. Our results establish the first robustness benchmarks for explanation agents and highlight robustness as a critical dimension for trustworthy, agent-driven recommender systems at web scale.

A DeepSeek-Powered AI System for Automated Chest Radiograph Interpretation in Clinical Practice

A global shortage of radiologists has been exacerbated by the significant volume of chest X-ray workloads, particularly in primary care. Although multimodal large language models show promise, existing evaluations predominantly rely on automated metrics or retrospective analyses, lacking rigorous prospective clinical validation. Janus-Pro-CXR (1B), a chest X-ray interpretation system based on DeepSeek Janus-Pro model, was developed and rigorously validated through a multicenter prospective trial (NCT07117266). Our system outperforms state-of-the-art X-ray report generation models in automated report generation, surpassing even larger-scale models including ChatGPT 4o (200B parameters), while demonstrating reliable detection of six clinically critical radiographic findings. Retrospective evaluation confirms significantly higher report accuracy than Janus-Pro and ChatGPT 4o. In prospective clinical deployment, AI assistance significantly improved report quality scores, reduced interpretation time by 18.3% (P < 0.001), and was preferred by a majority of experts in 54.3% of cases. Through lightweight architecture and domain-specific optimization, Janus-Pro-CXR improves diagnostic reliability and workflow efficiency, particularly in resource-constrained settings. The model architecture and implementation framework will be open-sourced to facilitate the clinical translation of AI-assisted radiology solutions.

Adaptive GPU Resource Allocation for Multi-Agent Collaborative Reasoning in Serverless Environments

Multi-agent systems powered by large language models have emerged as a promising paradigm for solving complex reasoning tasks through collaborative intelligence. However, efficiently deploying these systems on serverless GPU platforms presents significant resource allocation challenges due to heterogeneous agent workloads, varying computational demands, and the need for cost-effective scaling. This paper presents an adaptive GPU resource allocation framework that achieves 85\% latency reduction compared to round-robin scheduling while maintaining comparable throughput to static allocation, using an $O(N)$ complexity algorithm for real-time adaptation. Our approach dynamically allocates GPU resources based on workload characteristics, agent priorities, and minimum resource requirements, enabling efficient utilization while maintaining quality of service. The framework addresses three key challenges: (1) heterogeneous computational demands across lightweight coordinators and heavyweight specialists, (2) dynamic workload fluctuations requiring millisecond-scale reallocation, and (3) capacity constraints in serverless environments. Through comprehensive simulations modeling realistic multi-agent workflows with four heterogeneous agents, we demonstrate that adaptive allocation outperforms static equal and round-robin strategies across latency, cost, and GPU utilization metrics. The framework provides a practical solution for deploying cost-efficient multi-agent AI systems on serverless GPU infrastructure.