Intelligent agents designed for interactive environments face significant challenges in text-based games, a domain that demands complex reasoning and adaptability. While agents based on large language models (LLMs) using self-reflection have shown promise, they struggle when initially successful and exhibit reduced effectiveness when using smaller LLMs. We introduce Sweet&Sour, a novel approach that addresses these limitations in existing reflection methods by incorporating positive experiences and managed memory to enrich the context available to the agent at decision time. Our comprehensive analysis spans both closed- and open-source LLMs and demonstrates the effectiveness of Sweet&Sour in improving agent performance, particularly in scenarios where previous approaches fall short.

A Multi Robot System (MRS) is the infrastructure of an intelligent cyberphysical system, where the robots understand the need of the human, and hence cooperate together to fulfill this need. Modeling an MRS is a crucial aspect of designing the proper system architecture, because this model can be used to simulate and measure the performance of the proposed architecture. However, an MRS solution architecture modeling is a very difficult problem, as it contains many dependent behaviors that dynamically change due to the current status of the overall system. In this paper, we introduce a general purpose MRS case study, where the humans initiate requests that are achieved by the available robots. These requests require different plans that use the current capabilities of the available robots. After proposing an architecture that defines the solution components, three steps are followed. First is modeling these components via Business Process Model and Notation (BPMN) language. BPMN provides a graphical notation to precisely represent the behaviors of every component, which is an essential need to model the solution. Second is to simulate these components behaviors and interaction in form of software agents. Java Agent DEvelopment (JADE) middleware has been used to develop and simulate the proposed model. JADE is based on a reactive agent approach, therefore it can dynamically represent the interaction among the solution components. Finally is to analyze the performance of the solution by defining a number of quantitative measurements, which can be obtained while simulating the system model in JADE middleware, therefore the solution can be analyzed and compared to another architecture.

We introduce Vocal Sandbox, a framework for enabling seamless human-robot collaboration in situated environments. Systems in our framework are characterized by their ability to adapt and continually learn at multiple levels of abstraction from diverse teaching modalities such as spoken dialogue, object keypoints, and kinesthetic demonstrations. To enable such adaptation, we design lightweight and interpretable learning algorithms that allow users to build an understanding and co-adapt to a robot's capabilities in real-time, as they teach new behaviors. For example, after demonstrating a new low-level skill for "tracking around" an object, users are provided with trajectory visualizations of the robot's intended motion when asked to track a new object. Similarly, users teach high-level planning behaviors through spoken dialogue, using pretrained language models to synthesize behaviors such as "packing an object away" as compositions of low-level skills $-$ concepts that can be reused and built upon. We evaluate Vocal Sandbox in two settings: collaborative gift bag assembly and LEGO stop-motion animation. In the first setting, we run systematic ablations and user studies with 8 non-expert participants, highlighting the impact of multi-level teaching. Across 23 hours of total robot interaction time, users teach 17 new high-level behaviors with an average of 16 novel low-level skills, requiring 22.1% less active supervision compared to baselines and yielding more complex autonomous performance (+19.7%) with fewer failures (-67.1%). Qualitatively, users strongly prefer Vocal Sandbox systems due to their ease of use (+20.6%) and overall performance (+13.9%). Finally, we pair an experienced system-user with a robot to film a stop-motion animation; over two hours of continuous collaboration, the user teaches progressively more complex motion skills to shoot a 52 second (232 frame) movie.

Customer Relationship Management (CRM) systems are vital for modern enterprises, providing a foundation for managing customer interactions and data. Integrating AI agents into CRM systems can automate routine processes and enhance personalized service. However, deploying and evaluating these agents is challenging due to the lack of realistic benchmarks that reflect the complexity of real-world CRM tasks. To address this issue, we introduce CRMArena, a novel benchmark designed to evaluate AI agents on realistic tasks grounded in professional work environments. Following guidance from CRM experts and industry best practices, we designed CRMArena with nine customer service tasks distributed across three personas: service agent, analyst, and manager. The benchmark includes 16 commonly used industrial objects (e.g., account, order, knowledge article, case) with high interconnectivity, along with latent variables (e.g., complaint habits, policy violations) to simulate realistic data distributions. Experimental results reveal that state-of-the-art LLM agents succeed in less than 40% of the tasks with ReAct prompting, and less than 55% even with function-calling abilities. Our findings highlight the need for enhanced agent capabilities in function-calling and rule-following to be deployed in real-world work environments. CRMArena is an open challenge to the community: systems that can reliably complete tasks showcase direct business value in a popular work environment.

Autonomous agents powered by large vision and language models (VLM) have demonstrated significant potential in completing daily computer tasks, such as browsing the web to book travel and operating desktop software, which requires agents to understand these interfaces. Despite such visual inputs becoming more integrated into agentic applications, what types of risks and attacks exist around them still remain unclear. In this work, we demonstrate that VLM agents can be easily attacked by a set of carefully designed adversarial pop-ups, which human users would typically recognize and ignore. This distraction leads agents to click these pop-ups instead of performing the tasks as usual. Integrating these pop-ups into existing agent testing environments like OSWorld and VisualWebArena leads to an attack success rate (the frequency of the agent clicking the pop-ups) of 86% on average and decreases the task success rate by 47%. Basic defense techniques such as asking the agent to ignore pop-ups or including an advertisement notice, are ineffective against the attack.

This work proposes a novel set of techniques for approximating a Nash equilibrium in a finite, normal-form game. It achieves this by constructing a new reformulation as solving a parameterized system of multivariate polynomials with tunable complexity. In doing so, it forges an itinerant loop from game theory to machine learning and back. We show a Nash equilibrium can be approximated with purely calls to stochastic, iterative variants of singular value decomposition and power iteration, with implications for biological plausibility. We provide pseudocode and experiments demonstrating solving for all equilibria of a general-sum game using only these readily available linear algebra tools.

We study fair allocation of constrained resources, where a market designer optimizes overall welfare while maintaining group fairness. In many large-scale settings, utilities are not known in advance, but are instead observed after realizing the allocation. We therefore estimate agent utilities using machine learning. Optimizing over estimates requires trading-off between mean utilities and their predictive variances. We discuss these trade-offs under two paradigms for preference modeling -- in the stochastic optimization regime, the market designer has access to a probability distribution over utilities, and in the robust optimization regime they have access to an uncertainty set containing the true utilities with high probability. We discuss utilitarian and egalitarian welfare objectives, and we explore how to optimize for them under stochastic and robust paradigms. We demonstrate the efficacy of our approaches on three publicly available conference reviewer assignment datasets. The approaches presented enable scalable constrained resource allocation under uncertainty for many combinations of objectives and preference models.

Dynamic dispatching rules that allocate resources to tasks in real-time play a critical role in ensuring efficient operations of many automated material handling systems across industries. Traditionally, the dispatching rules deployed are typically the result of manually crafted heuristics based on domain experts' knowledge. Generating these rules is time-consuming and often sub-optimal. As enterprises increasingly accumulate vast amounts of operational data, there is significant potential to leverage this big data to enhance the performance of automated systems. One promising approach is to use Decision Transformers, which can be trained on existing enterprise data to learn better dynamic dispatching rules for improving system throughput. In this work, we study the application of Decision Transformers as dynamic dispatching policies within an actual multi-agent material handling system and identify scenarios where enterprises can effectively leverage Decision Transformers on existing big data to gain business value. Our empirical results demonstrate that Decision Transformers can improve the material handling system's throughput by a considerable amount when the heuristic originally used in the enterprise data exhibits moderate performance and involves no randomness. When the original heuristic has strong performance, Decision Transformers can still improve the throughput but with a smaller improvement margin. However, when the original heuristics contain an element of randomness or when the performance of the dataset is below a certain threshold, Decision Transformers fail to outperform the original heuristic. These results highlight both the potential and limitations of Decision Transformers as dispatching policies for automated industrial material handling systems.

Developing agents for complex and underspecified tasks, where no clear objective exists, remains challenging but offers many opportunities. This is especially true in video games, where simulated players (bots) need to play realistically, and there is no clear reward to evaluate them. While imitation learning has shown promise in such domains, these methods often fail when agents encounter out-of-distribution scenarios during deployment. Expanding the training dataset is a common solution, but it becomes impractical or costly when relying on human demonstrations. This article addresses active imitation learning, aiming to trigger expert intervention only when necessary, reducing the need for constant expert input along training. We introduce Random Network Distillation DAgger (RND-DAgger), a new active imitation learning method that limits expert querying by using a learned state-based out-of-distribution measure to trigger interventions. This approach avoids frequent expert-agent action comparisons, thus making the expert intervene only when it is useful. We evaluate RND-DAgger against traditional imitation learning and other active approaches in 3D video games (racing and third-person navigation) and in a robotic locomotion task and show that RND-DAgger surpasses previous methods by reducing expert queries. Imitation learning has increasingly become a favored approach for learning behaviors in complex environments, offering a compelling alternative to classical scripted behaviors implemented by domain specialists (Schaal, 1999; Hussein et al., 2017). It is particularly well suited in problems where there is not a clear performance measure (or reward).

We present a framework for Multi-Robot Task Allocation (MRTA) in heterogeneous teams performing long-endurance missions in dynamic scenarios. Given the limited battery of robots, especially in the case of aerial vehicles, we allow for robot recharges and the possibility of fragmenting and/or relaying certain tasks. We also address tasks that must be performed by a coalition of robots in a coordinated manner. Given these features, we introduce a new class of heterogeneous MRTA problems which we analyze theoretically and optimally formulate as a Mixed-Integer Linear Program. We then contribute a heuristic algorithm to compute approximate solutions and integrate it into a mission planning and execution architecture capable of reacting to unexpected events by repairing or recomputing plans online. Our experimental results show the relevance of our newly formulated problem in a realistic use case for inspection with aerial robots. We assess the performance of our heuristic solver in comparison with other variants and with exact optimal solutions in small-scale scenarios. In addition, we evaluate the ability of our replanning framework to repair plans online.