Gonzalo Gonzalez-Pumariega

GameEngineBench: Evaluating Coding Agents on Real C++ Runtime Environments

Brian La, Sejoon Chang, Ben Kim, Jihye Bae et autres

Game engines provide real-time simulation, rendering, physics, interaction, networking, and asset pipelines, making them valuable not only for games but also for 3D applications in healthcare, robotics, architecture, manufacturing, and related domains. Because game development is where these systems are most mature …

GameEngineBench: Evaluating Coding Agents on Real C++ Runtime Environments

Brian La, Sejoon Chang, Ben Kim, Jihye Bae et autres

Game engines provide real-time simulation, rendering, physics, interaction, networking, and asset pipelines, making them valuable not only for games but also for 3D applications in healthcare, robotics, architecture, manufacturing, and related domains. Because game development is where these systems are most mature …

us (code pays fourni par la source)

On the Reliability of Computer Use Agents

Gonzalo Gonzalez-Pumariega, Saaket Agashe, Jiachen Yang, A. Li et autres

Computer-use agents have rapidly improved on real-world tasks such as web navigation, desktop automation, and software interaction, in some cases surpassing human performance. Yet even when the task and model are unchanged, an agent that succeeds once may fail on a repeated …

On the Reliability of Computer Use Agents

Gonzalo Gonzalez-Pumariega, Saaket Agashe, Jiachen Yang, Ang Li et autres

Computer-use agents have rapidly improved on real-world tasks such as web navigation, desktop automation, and software interaction, in some cases surpassing human performance. Yet even when the task and model are unchanged, an agent that succeeds once may fail on a repeated …

Scaling Agents for Computer Use

Gonzalo Gonzalez-Pumariega, Vincent Tu, Jun-Yao Yang, Ang Li et autres

Computer-use agents (CUAs) hold promise for automating everyday digital tasks, but their performance on long-horizon, complex problems remains unreliable. Single-rollout execution is brittle, with small errors compounding over time and leading to high variance in outcomes. While prior work has attempted to …

Multi-Turn Code Generation Through Single-Step Rewards

Arnav Jain, Gonzalo Gonzalez-Pumariega, Wayne Chen, Alexander M. Rush et autres

We address the problem of code generation from multi-turn execution feedback. Existing methods either generate code without feedback or use complex, hierarchical reinforcement learning to optimize multi-turn rewards. We propose a simple yet scalable approach, $μ$Code, that solves multi-turn code generation using …

Robotouille: An Asynchronous Planning Benchmark for LLM Agents

Gonzalo Gonzalez-Pumariega, Leong Su Yean, Neha Sunkara, Sanjiban Choudhury

Effective asynchronous planning, or the ability to efficiently reason and plan over states and actions that must happen in parallel or sequentially, is essential for agents that must account for time delays, reason over diverse long-horizon tasks, and collaborate with other agents. …

Query-Efficient Planning with Language Models

Gonzalo Gonzalez-Pumariega, Wayne Chen, Kushal S. Kedia, Sanjiban Choudhury

Planning in complex environments requires an agent to efficiently query a world model to find a feasible sequence of actions from start to goal. Recent work has shown that Large Language Models (LLMs), with their rich prior knowledge and reasoning capabilities, can …

APRICOT: Active Preference Learning and Constraint-Aware Task Planning with LLMs

Huaxiaoyue Wang, Nathaniel A. Chin, Gonzalo Gonzalez-Pumariega, Neha Sunkara et autres

Home robots performing personalized tasks must adeptly balance user preferences with environmental affordances. We focus on organization tasks within constrained spaces, such as arranging items into a refrigerator, where preferences for placement collide with physical limitations. The robot must infer user preferences …

MOSAIC: Modular Foundation Models for Assistive and Interactive Cooking

Huaxiaoyue Wang, Kushal S. Kedia, Juntao Ren, Atiksh Bhardwaj et autres

We present MOSAIC, a modular architecture for coordinating multiple robots to (a) interact with users using natural language and (b) manipulate an open vocabulary of everyday objects. MOSAIC employs modularity at several levels: it leverages multiple large-scale pre-trained models for high-level tasks …

Affinity Diagramming with a Robot

Matthew V. Law, Nnamdi Nwagwu, Amritansh Kwatra, Seo-Young Lee et autres

We investigate what it might look like for a robot to work with a human on a need-finding design task using an affinity diagram. While some recent projects have examined how human–robot teams might explore solutions to design problems, human–robot collaboration in …

us (code pays fourni par la source)

Urgency Builds Trust: A Voice Agent's Emotional Expression in an Emergency

Jieun Kim, Gonzalo Gonzalez-Pumariega, Soyee Park, Susan R. Fussell

Voice agents play a crucial role in providing verbal guidance to users, especially during emergency situations to ensure their successful evacuation. While the vocal presentation of emotion significantly impacts communication effectiveness in human-agent interaction, the context-dependent impacts of emotional voice in an …

us (code pays fourni par la source)