In this work, we study the risks of collective financial fraud in large-scale multi-agent systems powered by large language model (LLM) agents. We investigate whether agents can collaborate in fraudulent behaviors, how such collaboration amplifies risks, and what factors influence fraud success. To support this research, we present MultiAgentFraudBench, a large-scale benchmark for simulating financial fraud scenarios based on realistic online interactions. The benchmark covers 28 typical online fraud scenarios, spanning the full fraud lifecycle across both public and private domains. We further analyze key factors affecting fraud success, including interaction depth, activity level, and fine-grained collaboration failure modes. Finally, we propose a series of mitigation strategies, including adding content-level warnings to fraudulent posts and dialogues, using LLMs as monitors to block potentially malicious agents, and fostering group resilience through information sharing at the societal level. Notably, we observe that malicious agents can adapt to environmental interventions. Our findings highlight the real-world risks of multi-agent financial fraud and suggest practical measures for mitigating them. Code is available at https://github.com/zheng977/MutiAgent4Fraud.

As AI moves beyond text, large language models (LLMs) increasingly power vision, audio, and document understanding; however, their high inference costs hinder real-time, scalable deployment. Conversely, smaller open-source models offer cost advantages but struggle with complex or multimodal queries. We introduce a unified, modular framework that intelligently routes each query - textual, multimodal, or complex - to the most fitting expert model, using a learned routing network that balances cost and quality. For vision tasks, we employ a two-stage open-source pipeline optimized for efficiency and reviving efficient classical vision components where they remain SOTA for sub-tasks. On benchmarks such as Massive Multitask Language Understanding (MMLU) and Visual Question Answering (VQA), we match or exceed the performance of always-premium LLM (monolithic systems with one model serving all query types) performance, yet reduce the reliance on costly models by over 67%. With its extensible, multi-agent orchestration, we deliver high-quality, resource-efficient AI at scale.

Recent advancements in GUI agents have significantly expanded their ability to interpret natural language commands to manage software interfaces. However, acquiring GUI data remains a significant challenge. Existing methods often involve designing automated agents that browse URLs from the Common Crawl, using webpage HTML to collect screenshots and corresponding annotations, including the names and bounding boxes of UI elements. However, this method is difficult to apply to desktop software or some newly launched websites not included in the Common Crawl. While we expect the model to possess strong generalization capabilities to handle this, it is still crucial for personalized scenarios that require rapid and perfect adaptation to new software or websites. To address this, we propose an automated data collection method with minimal annotation costs, named Auto-Explorer. It incorporates a simple yet effective exploration mechanism that autonomously parses and explores GUI environments, gathering data efficiently. Additionally, to assess the quality of exploration, we have developed the UIXplore benchmark. This benchmark creates environments for explorer agents to discover and save software states. Using the data gathered, we fine-tune a multimodal large language model (MLLM) and establish a GUI element grounding testing set to evaluate the effectiveness of the exploration strategies. Our experiments demonstrate the superior performance of Auto-Explorer, showing that our method can quickly enhance the capabilities of an MLLM in explored software.

Safety evaluation of large language models (LLMs) increasingly relies on LLM-as-a-Judge frameworks, but the high cost of frontier models limits scalability. We propose a cost-efficient multi-agent judging framework that employs Small Language Models (SLMs) through structured debates among critic, defender, and judge agents. To rigorously assess safety judgments, we construct HAJailBench, a large-scale human-annotated jailbreak benchmark comprising 12,000 adversarial interactions across diverse attack methods and target models. The dataset provides fine-grained, expert-labeled ground truth for evaluating both safety robustness and judge reliability. Our SLM-based framework achieves agreement comparable to GPT-4o judges on HAJailBench while substantially reducing inference cost. Ablation results show that three rounds of debate yield the optimal balance between accuracy and efficiency. These findings demonstrate that structured, value-aligned debate enables SLMs to capture semantic nuances of jailbreak attacks and that HAJailBench offers a reliable foundation for scalable LLM safety evaluation.

Organizations are increasingly exploring delegation of screening and negotiation tasks to AI systems, yet deployment in high-stakes B2B settings is constrained by governance: preventing unauthorized commitments, ensuring sufficient information before bargaining, and maintaining effective human oversight and auditability. Prior work on large language model negotiation largely emphasizes autonomous bargaining between agents and omits practical needs such as staged information gathering, explicit authorization boundaries, and systematic feedback integration. We propose GAIA, a governance-first framework for LLM-human agency in B2B negotiation and screening. GAIA defines three essential roles - Principal (human), Delegate (LLM agent), and Counterparty - with an optional Critic to enhance performance, and organizes interactions through three mechanisms: information-gated progression that separates screening from negotiation; dual feedback integration that combines AI critique with lightweight human corrections; and authorization boundaries with explicit escalation paths. Our contributions are fourfold: (1) a formal governance framework with three coordinated mechanisms and four safety invariants for delegation with bounded authorization; (2) information-gated progression via task-completeness tracking (TCI) and explicit state transitions that separate screening from commitment; (3) dual feedback integration that blends Critic suggestions with human oversight through parallel learning channels; and (4) a hybrid validation blueprint that combines automated protocol metrics with human judgment of outcomes and safety. By bridging theory and practice, GAIA offers a reproducible specification for safe, efficient, and accountable AI delegation that can be instantiated across procurement, real estate, and staffing workflows.

We introduce MemoriesDB, a unified data architecture designed to avoid decoherence across time, meaning, and relation in long-term computational memory. Each memory is a time-semantic-relational entity-a structure that simultaneously encodes when an event occurred, what it means, and how it connects to other events. Built initially atop PostgreSQL with pgvector extensions, MemoriesDB combines the properties of a time-series datastore, a vector database, and a graph system within a single append-only schema. Each memory is represented as a vertex uniquely labeled by its microsecond timestamp and accompanied by low- and high-dimensional normalized embeddings that capture semantic context. Directed edges between memories form labeled relations with per-edge metadata, enabling multiple contextual links between the same vertices. Together these constructs form a time-indexed stack of temporal-semantic surfaces, where edges project as directional arrows in a 1+1-dimensional similarity field, tracing the evolution of meaning through time while maintaining cross-temporal coherence. This formulation supports efficient time-bounded retrieval, hybrid semantic search, and lightweight structural reasoning in a single query path. A working prototype demonstrates scalable recall and contextual reinforcement using standard relational infrastructure, and we discuss extensions toward a columnar backend, distributed clustering, and emergent topic modeling.

Monte Carlo Tree Search (MCTS), which leverages Upper Confidence Bound for Trees (UCTs) to balance exploration and exploitation through randomized sampling, is instrumental to solving complex planning problems. However, for multi-agent planning, MCTS is confronted with a large combinatorial action space that often grows exponentially with the number of agents. As a result, the branching factor of MCTS during tree expansion also increases exponentially, making it very difficult to efficiently explore and exploit during tree search. To this end, we propose MALinZero, a new approach to leverage low-dimensional representational structures on joint-action returns and enable efficient MCTS in complex multi-agent planning. Our solution can be viewed as projecting the joint-action returns into the low-dimensional space representable using a contextual linear bandit problem formulation. We solve the contextual linear bandit problem with convex and $μ$-smooth loss functions -- in order to place more importance on better joint actions and mitigate potential representational limitations -- and derive a linear Upper Confidence Bound applied to trees (LinUCT) to enable novel multi-agent exploration and exploitation in the low-dimensional space. We analyze the regret of MALinZero for low-dimensional reward functions and propose an $(1-\tfrac1e)$-approximation algorithm for the joint action selection by maximizing a sub-modular objective. MALinZero demonstrates state-of-the-art performance on multi-agent benchmarks such as matrix games, SMAC, and SMACv2, outperforming both model-based and model-free multi-agent reinforcement learning baselines with faster learning speed and better performance.

The current body of research on Parkinson's disease (PD) screening, monitoring, and management has evolved along two largely independent trajectories. The first research community focuses on multimodal sensing of PD-related biomarkers using noninvasive technologies such as inertial measurement units (IMUs), force/pressure insoles, electromyography (EMG), electroencephalography (EEG), speech and acoustic analysis, and RGB/RGB-D motion capture systems. These studies emphasize data acquisition, feature extraction, and machine learning-based classification for PD screening, diagnosis, and disease progression modeling. In parallel, a second research community has concentrated on robotic intervention and rehabilitation, employing socially assistive robots (SARs), robot-assisted rehabilitation (RAR) systems, and virtual reality (VR)-integrated robotic platforms for improving motor and cognitive function, enhancing social engagement, and supporting caregivers. Despite the complementary goals of these two domains, their methodological and technological integration remains limited, with minimal data-level or decision-level coupling between the two. With the advent of advanced artificial intelligence (AI), including large language models (LLMs), agentic AI systems, a unique opportunity now exists to unify these research streams. We envision a closed-loop sensor-AI-robot framework in which multimodal sensing continuously guides the interaction between the patient, caregiver, humanoid robot (and physician) through AI agents that are powered by a multitude of AI models such as robotic and wearables foundation models, LLM-based reasoning, reinforcement learning, and continual learning. Such closed-loop system enables personalized, explainable, and context-aware intervention, forming the basis for digital twin of the PD patient that can adapt over time to deliver intelligent, patient-centered PD care.

While reinforcement learning (RL) can empower autonomous agents by enabling self-improvement through interaction, its practical adoption remains challenging due to costly rollouts, limited task diversity, unreliable reward signals, and infrastructure complexity, all of which obstruct the collection of scalable experience data. To address these challenges, we introduce DreamGym, the first unified framework designed to synthesize diverse experiences with scalability in mind to enable effective online RL training for autonomous agents. Rather than relying on expensive real-environment rollouts, DreamGym distills environment dynamics into a reasoning-based experience model that derives consistent state transitions and feedback signals through step-by-step reasoning, enabling scalable agent rollout collection for RL. To improve the stability and quality of transitions, DreamGym leverages an experience replay buffer initialized with offline real-world data and continuously enriched with fresh interactions to actively support agent training. To improve knowledge acquisition, DreamGym adaptively generates new tasks that challenge the current agent policy, enabling more effective online curriculum learning. Experiments across diverse environments and agent backbones demonstrate that DreamGym substantially improves RL training, both in fully synthetic settings and in sim-to-real transfer scenarios. On non-RL-ready tasks like WebArena, DreamGym outperforms all baselines by over 30%. And in RL-ready but costly settings, it matches GRPO and PPO performance using only synthetic interactions. When transferring a policy trained purely on synthetic experiences to real-environment RL, DreamGym yields significant additional performance gains while requiring far fewer real-world interactions, providing a scalable warm-start strategy for general-purpose RL.

Hyperchat AI is a novel agentic technology that enables thoughtful conversations among networked human groups of potentially unlimited size. It allows large teams to discuss complex issues, brainstorm ideas, surface risks, assess alternatives and efficiently converge on optimized solutions that amplify the group's Collective Intelligence (CI). A formal study was conducted to quantify the forecasting accuracy of human groups using Hyperchat AI to conversationally predict the outcome of Major League Baseball (MLB) games. During an 8-week period, networked groups of approximately 24 sports fans were tasked with collaboratively forecasting the winners of 59 baseball games through real-time conversation facilitated by AI agents. The results showed that when debating the games using Hyperchat AI technology, the groups converged on High Confidence predictions that significantly outperformed Vegas betting markets. Specifically, groups were 78% accurate in their High Confidence picks, a statistically strong result vs the Vegas odds of 57% (p=0.020). Had the groups bet against the spread (ATS) on these games, they would have achieved a 46% ROI against Vegas betting markets. In addition, High Confidence forecasts that were generated through above-average conversation rates were 88% accurate, suggesting that real-time interactive deliberation is central to amplified accuracy.