114 citations · 277 across the 33 of their papers we have counts for
34 papers
The Gaussian Is Enough: Flow-Matching Priors Do Not Help When Fine-Tuning Large Behavior Models
Chen Xu, Rishi Shah, Hadas Kress-Gazit +2
Modern robot imitation learning increasingly relies on generative policies based on diffusion or flow-matching models, which generate actions by transforming samples from a prior d…
Impact of Different Failures on a Robot's Perceived Reliability
Andrew Violette, Zhanxin Wu, Haruki Nishimura +5
Robots fail, potentially leading to a loss in the robot's perceived reliability (PR), a measure correlated with trustworthiness. In this study we examine how various kinds of failu…
Physically-Feasible Reactive Synthesis for Terrain-Adaptive Locomotion
Ziyi Zhou, Qian Meng, Hadas Kress-Gazit +1
We present an integrated planning framework for quadrupedal locomotion over dynamically changing, unforeseen terrains. Existing methods often depend on heuristics for real-time foo…
A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation
TRI LBM Team, Jose Barreiros, Andrew Beaulieu +79
Robot manipulation has seen tremendous progress in recent years, with imitation learning policies enabling successful performance of dexterous and hard-to-model tasks. Concurrently…
Report on NSF Workshop on Science of Safe AI
Rajeev Alur, Greg Durrett, Hadas Kress-Gazit +2
Recent advances in machine learning, particularly the emergence of foundation models, are leading to new opportunities to develop technology-based solutions to societal problems. H…
Physically-Feasible Reactive Synthesis for Terrain-Adaptive Locomotion via Trajectory Optimization and Symbolic Repair
Ziyi Zhou, Qian Meng, Hadas Kress-Gazit +1
We propose an integrated planning framework for quadrupedal locomotion over dynamically changing, unforeseen terrains. Existing approaches either rely on heuristics for instantaneo…
INPROVF: Leveraging Large Language Models to Repair High-level Robot Controllers from Assumption Violations
Qian Meng, Jin Peng Zhou, Kilian Q. Weinberger +1
This paper presents INPROVF, an automatic framework that combines large language models (LLMs) and formal methods to speed up the repair process of high-level robot controllers. Pr…
Robot Learning as an Empirical Science: Best Practices for Policy Evaluation
Hadas Kress-Gazit, Kunimatsu Hashimoto, Naveen Kuppuswamy +5
The robot learning community has made great strides in recent years, proposing new architectures and showcasing impressive new capabilities; however, the dominant metric used in th…
Online Resynthesis of High-Level Collaborative Tasks for Robots with Changing Capabilities
Amy Fang, Tenny Yin, Hadas Kress-Gazit
Given a collaborative high-level task and a team of heterogeneous robots and behaviors to satisfy it, this work focuses on the challenge of automatically, at runtime, adjusting the…
Automated Robot Recovery from Assumption Violations of High-Level Specifications
Qian Meng, Hadas Kress-Gazit
This paper presents a framework that enables robots to automatically recover from assumption violations of high-level specifications during task execution. In contrast to previous…
Continuous Execution of High-Level Collaborative Tasks for Heterogeneous Robot Teams
Amy Fang, Tenny Yin, Jiawei Lin +1
We propose a control synthesis framework for a heterogeneous multi-robot system to satisfy collaborative tasks, where actions may take varying duration of time to complete. We enco…
Multi-Source Encapsulation With Guaranteed Convergence Using Minimalist Robots
Himani Sinhmar, Hadas Kress-Gazit
We present a decentralized control algorithm for a minimalist robotic swarm lacking memory, explicit communication, or relative position information, to encapsulate multiple diffus…
High-Level, Collaborative Task Planning Grammar and Execution for Heterogeneous Agents
Amy Fang, Hadas Kress-Gazit
We propose a new multi-agent task grammar to encode collaborative tasks for a team of heterogeneous agents that can have overlapping capabilities. The grammar allows users to speci…
Ensuring Reliable Robot Task Performance through Probabilistic Rare-Event Verification and Synthesis
Guy Scher, Sadra Sadraddini, Ariel Yadin +1
Providing guarantees on the safe operation of robots against edge cases is challenging as testing methods such as traditional Monte-Carlo require too many samples to provide reason…
Online Modifications for Event-based Signal Temporal Logic Specifications
David Gundana, Hadas Kress-Gazit
In this paper we present a grammar and control synthesis framework for online modification of Event-based Signal Temporal Logic (STL) specifications, during execution. These modifi…
Guaranteed Encapsulation of Targets with Unknown Motion by a Minimalist Robotic Swarm
Himani Sinhmar, Hadas Kress-Gazit
We present a decentralized control algorithm for a robotic swarm given the task of encapsulating static and moving targets in a bounded unknown environment. We consider minimalist…
Decentralized Control of Minimalistic Robotic Swarms For Guaranteed Target Encapsulation
Himani Sinhmar, Hadas Kress-Gazit
We propose a decentralized control algorithm for a minimalistic robotic swarm with limited capabilities such that the desired global behavior emerges. We consider the problem of se…
Physically-Feasible Repair of Reactive, Linear Temporal Logic-based, High-Level Tasks
Adam Pacheck, Hadas Kress-Gazit
A typical approach to creating complex robot behaviors is to compose atomic controllers, or skills, such that the resulting behavior satisfies a high-level task; however, when a ta…
Automatic Encoding and Repair of Reactive High-Level Tasks with Learned Abstract Representations
Adam Pacheck, Steven James, George Konidaris +1
We present a framework that, given a set of skills a robot can perform, abstracts sensor data into symbols that we use to automatically encode the robot's capabilities in Linear Te…
Automated Task Updates of Temporal Logic Specifications for Heterogeneous Robots
Amy Fang, Hadas Kress-Gazit
Given a heterogeneous group of robots executing a complex task represented in Linear Temporal Logic, and a new set of tasks for the group, we define the task update problem and pro…
Elliptical Slice Sampling for Probabilistic Verification of Stochastic Systems with Signal Temporal Logic Specifications
Guy Scher, Sadra Sadraddini, Russ Tedrake +1
Autonomous robots typically incorporate complex sensors in their decision-making and control loops. These sensors, such as cameras and Lidars, have imperfections in their sensing a…
Synthesizing Modular Manipulators For Tasks With Time, Obstacle, And Torque Constraints
Thais Campos, Hadas Kress-Gazit
Modular robots can be tailored to achieve specific tasks and rearranged to achieve previously infeasible ones. The challenge is choosing an appropriate design from a large search s…
Counterexample-Guided Repair for Symbolic-Geometric Action Abstractions
Wil Thomason, Hadas Kress-Gazit
Integrated Task and Motion Planning (TMP) provides a promising class of approaches for solving robot planning problems with intricate symbolic and geometric constraints. However, t…
Learning and Planning for Temporally Extended Tasks in Unknown Environments
Christopher Bradley, Adam Pacheck, Gregory J. Stein +3
We propose a novel planning technique for satisfying tasks specified in temporal logic in partially revealed environments. We define high-level actions derived from the environment…