Search Results for author: Sanjit A. Seshia

Found 48 papers, 12 papers with code

$L^*LM$: Learning Automata from Examples using Natural Language Oracles

no code implementations10 Feb 2024 Marcell Vazquez-Chanlatte, Karim Elmaaroufi, Stefan J. Witwicki, Sanjit A. Seshia

Due to the expressivity of natural language, we observe a significant improvement in the data efficiency of learning DFAs from expert demonstrations.

An Eager Satisfiability Modulo Theories Solver for Algebraic Datatypes

no code implementations18 Oct 2023 Amar Shah, Federico Mora, Sanjit A. Seshia

Specifically, our solver reduces ADT queries to a simpler logical theory, uninterpreted functions (UF), and then uses an existing solver on the reduced query.

Specification-Guided Data Aggregation for Semantically Aware Imitation Learning

no code implementations29 Mar 2023 Ameesh Shah, Jonathan DeCastro, John Gideon, Beyazit Yalcinkaya, Guy Rosman, Sanjit A. Seshia

Advancements in simulation and formal methods-guided environment sampling have enabled the rigorous evaluation of machine learning models in a number of safety-critical scenarios, such as autonomous driving.

Autonomous Driving Imitation Learning

Demonstration Informed Specification Search

1 code implementation20 Dec 2021 Marcell Vazquez-Chanlatte, Ameesh Shah, Gil Lederman, Sanjit A. Seshia

This paper considers the problem of learning temporal task specifications, e. g. automata and temporal logic, from expert demonstrations.

A Scenario-Based Platform for Testing Autonomous Vehicle Behavior Prediction Models in Simulation

no code implementations28 Oct 2021 Francis Indaheng, Edward Kim, Kesav Viswanadha, Jay Shenoy, Jinkyu Kim, Daniel J. Fremont, Sanjit A. Seshia

Hence, it is important that these prediction models are extensively tested in various test scenarios involving interactive behaviors prior to deployment.

Probabilistic Programming

Synthesizing Pareto-Optimal Interpretations for Black-Box Models

no code implementations16 Aug 2021 Hazem Torfah, Shetal Shah, Supratik Chakraborty, S. Akshay, Sanjit A. Seshia

For a given black-box, our approach yields a set of Pareto-optimal interpretations with respect to the correctness and explainability measures.

Satisfiability and Synthesis Modulo Oracles

no code implementations28 Jul 2021 Elizabeth Polgreen, Andrew Reynolds, Sanjit A. Seshia

As a necessary component of this framework, we also formalize the problem of satisfiability modulo theories and oracles, and present an algorithm for solving this problem.

Program Synthesis

Parallel and Multi-Objective Falsification with Scenic and VerifAI

1 code implementation9 Jul 2021 Kesav Viswanadha, Edward Kim, Francis Indaheng, Daniel J. Fremont, Sanjit A. Seshia

Falsification has emerged as an important tool for simulation-based verification of autonomous systems.

Scenic4RL: Programmatic Modeling and Generation of Reinforcement Learning Environments

no code implementations18 Jun 2021 Abdus Salam Azad, Edward Kim, Qiancheng Wu, Kimin Lee, Ion Stoica, Pieter Abbeel, Sanjit A. Seshia

To showcase the benefits, we interfaced SCENIC to an existing RTS environment Google Research Football(GRF) simulator and introduced a benchmark consisting of 32 realistic scenarios, encoded in SCENIC, to train RL agents and testing their generalization capabilities.

reinforcement-learning Reinforcement Learning (RL)

Scenic: A Language for Scenario Specification and Data Generation

2 code implementations13 Oct 2020 Daniel J. Fremont, Edward Kim, Tommaso Dreossi, Shromona Ghosh, Xiangyu Yue, Alberto L. Sangiovanni-Vincentelli, Sanjit A. Seshia

We design a domain-specific language, Scenic, for describing scenarios that are distributions over scenes and the behaviors of their agents over time.

Probabilistic Programming Synthetic Data Generation

A Review of Single-Source Deep Unsupervised Visual Domain Adaptation

1 code implementation1 Sep 2020 Sicheng Zhao, Xiangyu Yue, Shanghang Zhang, Bo Li, Han Zhao, Bichen Wu, Ravi Krishna, Joseph E. Gonzalez, Alberto L. Sangiovanni-Vincentelli, Sanjit A. Seshia, Kurt Keutzer

To cope with limited labeled training data, many have attempted to directly apply models trained on a large-scale labeled source domain to another sparsely labeled or unlabeled target domain.

Unsupervised Domain Adaptation

SOTER on ROS: A Run-Time Assurance Framework on the Robot Operating System

1 code implementation21 Aug 2020 Sumukh Shivakumar, Hazem Torfah, Ankush Desai, Sanjit A. Seshia

We present an implementation of SOTER, a run-time assurance framework for building safe distributed mobile robotic (DMR) systems, on top of the Robot Operating System (ROS).

Gradient Descent over Metagrammars for Syntax-Guided Synthesis

no code implementations13 Jul 2020 Nicolas Chan, Elizabeth Polgreen, Sanjit A. Seshia

The performance of a syntax-guided synthesis algorithm is highly dependent on the provision of a good syntactic template, or grammar.

Learning Branching Heuristics for Propositional Model Counting

no code implementations7 Jul 2020 Pashootan Vaezipoor, Gil Lederman, Yuhuai Wu, Chris J. Maddison, Roger Grosse, Sanjit A. Seshia, Fahiem Bacchus

In addition to step count improvements, Neuro# can also achieve orders of magnitude wall-clock speedups over the vanilla solver on larger instances in some problem families, despite the runtime overhead of querying the model.

Enforcing Almost-Sure Reachability in POMDPs

1 code implementation30 Jun 2020 Sebastian Junges, Nils Jansen, Sanjit A. Seshia

Partially-Observable Markov Decision Processes (POMDPs) are a well-known stochastic model for sequential decision making under limited information.

Decision Making reinforcement-learning +2

Formal Analysis and Redesign of a Neural Network-Based Aircraft Taxiing System with VerifAI

no code implementations14 May 2020 Daniel J. Fremont, Johnathan Chiu, Dragos D. Margineantu, Denis Osipychev, Sanjit A. Seshia

We demonstrate a unified approach to rigorous design of safety-critical autonomous systems using the VerifAI toolkit for formal analysis of AI-based systems.

Probabilistic Programming

Formal Scenario-Based Testing of Autonomous Vehicles: From Simulation to the Real World

no code implementations17 Mar 2020 Daniel J. Fremont, Edward Kim, Yash Vardhan Pant, Sanjit A. Seshia, Atul Acharya, Xantha Bruso, Paul Wells, Steve Lemke, Qiang Lu, Shalin Mehta

We present a new approach to automated scenario-based testing of the safety of autonomous vehicles, especially those using advanced artificial intelligence-based components, spanning both simulation-based evaluation as well as testing in the real world.

Autonomous Vehicles

Modularity in Query-Based Concept Learning

no code implementations7 Nov 2019 Benjamin Caulfield, Sanjit A. Seshia

We define and study the problem of modular concept learning, that is, learning a concept that is a cross product of component concepts.

Counterexample-Guided Synthesis of Perception Models and Control

no code implementations4 Nov 2019 Shromona Ghosh, Yash Vardhan Pant, Hadi Ravanbakhsh, Sanjit A. Seshia

The framework uses a falsifier to find counterexamples, or traces of the systems that violate a safety property, to extract information that enables efficient modeling of the perception modules and errors in it.

Autonomous Vehicles

On the Utility of Learning about Humans for Human-AI Coordination

2 code implementations NeurIPS 2019 Micah Carroll, Rohin Shah, Mark K. Ho, Thomas L. Griffiths, Sanjit A. Seshia, Pieter Abbeel, Anca Dragan

While we would like agents that can coordinate with humans, current algorithms such as self-play and population-based training create agents that can coordinate with themselves.

Maximum Causal Entropy Specification Inference from Demonstrations

no code implementations26 Jul 2019 Marcell Vazquez-Chanlatte, Sanjit A. Seshia

In many settings (e. g., robotics) demonstrations provide a natural way to specify tasks; however, most methods for learning from demonstrations either do not provide guarantees that the artifacts learned for the tasks, such as rewards or policies, can be safely composed and/or do not explicitly capture history dependencies.

A Formalization of Robustness for Deep Neural Networks

no code implementations24 Mar 2019 Tommaso Dreossi, Shromona Ghosh, Alberto Sangiovanni-Vincentelli, Sanjit A. Seshia

The process of generating the perturbations that expose the lack of robustness of neural networks is known as adversarial input generation.

Adversarial Attack

A Model Counter's Guide to Probabilistic Systems

no code implementations22 Mar 2019 Marcell Vazquez-Chanlatte, Markus N. Rabe, Sanjit A. Seshia

In this paper, we systematize the modeling of probabilistic systems for the purpose of analyzing them with model counting techniques.

VERIFAI: A Toolkit for the Design and Analysis of Artificial Intelligence-Based Systems

1 code implementation12 Feb 2019 Tommaso Dreossi, Daniel J. Fremont, Shromona Ghosh, Edward Kim, Hadi Ravanbakhsh, Marcell Vazquez-Chanlatte, Sanjit A. Seshia

We present VERIFAI, a software toolkit for the formal design and analysis of systems that include artificial intelligence (AI) and machine learning (ML) components.

BIG-bench Machine Learning

Scenic: A Language for Scenario Specification and Scene Generation

2 code implementations25 Sep 2018 Daniel J. Fremont, Tommaso Dreossi, Shromona Ghosh, Xiangyu Yue, Alberto L. Sangiovanni-Vincentelli, Sanjit A. Seshia

We propose a new probabilistic programming language for the design and analysis of perception systems, especially those based on machine learning.

Probabilistic Programming Scene Generation +1

Cloud-based Quadratic Optimization with Partially Homomorphic Encryption

1 code implementation7 Sep 2018 Andreea B. Alexandru, Konstantinos Gatsis, Yasser Shoukry, Sanjit A. Seshia, Paulo Tabuada, George J. Pappas

The development of large-scale distributed control systems has led to the outsourcing of costly computations to cloud-computing platforms, as well as to concerns about privacy of the collected sensitive data.

Optimization and Control Cryptography and Security Systems and Control

SOTER: A Runtime Assurance Framework for Programming Safe Robotics Systems

no code implementations23 Aug 2018 Ankush Desai, Shromona Ghosh, Sanjit A. Seshia, Natarajan Shankar, Ashish Tiwari

SOTER provides language primitives to declaratively construct a RTA module consisting of an advanced, high-performance controller (uncertified), a safe, lower-performance controller (certified), and the desired safety specification.

Learning Heuristics for Quantified Boolean Formulas through Deep Reinforcement Learning

1 code implementation20 Jul 2018 Gil Lederman, Markus N. Rabe, Edward A. Lee, Sanjit A. Seshia

We demonstrate how to learn efficient heuristics for automated reasoning algorithms for quantified Boolean formulas through deep reinforcement learning.

reinforcement-learning Reinforcement Learning (RL)

Semantic Adversarial Deep Learning

no code implementations19 Apr 2018 Tommaso Dreossi, Somesh Jha, Sanjit A. Seshia

However, existing approaches to generating adversarial examples and devising robust ML algorithms mostly ignore the semantics and context of the overall system containing the ML component.

Malware Detection Self-Driving Cars

Unsupervised Domain Adaptation: from Simulation Engine to the RealWorld

no code implementations24 Mar 2018 Sicheng Zhao, Bichen Wu, Joseph Gonzalez, Sanjit A. Seshia, Kurt Keutzer

To cope with limited labeled training data, many have attempted to directly apply models trained on a large-scale labeled source domain to another sparsely labeled target domain.

Unsupervised Domain Adaptation

Context-Specific Validation of Data-Driven Models

no code implementations14 Feb 2018 Somil Bansal, Shromona Ghosh, Alberto Sangiovanni-Vincentelli, Sanjit A. Seshia, Claire J. Tomlin

We propose a context-specific validation framework to quantify the quality of a learned model based on a distance measure between the closed-loop actual system and the learned model.

Learning Task Specifications from Demonstrations

no code implementations NeurIPS 2018 Marcell Vazquez-Chanlatte, Susmit Jha, Ashish Tiwari, Mark K. Ho, Sanjit A. Seshia

In this paper, we formulate the specification inference task as a maximum a posteriori (MAP) probability inference problem, apply the principle of maximum entropy to derive an analytic demonstration likelihood model and give an efficient approach to search for the most likely specification in a large candidate pool of specifications.

Systematic Testing of Convolutional Neural Networks for Autonomous Driving

no code implementations10 Aug 2017 Tommaso Dreossi, Shromona Ghosh, Alberto Sangiovanni-Vincentelli, Sanjit A. Seshia

We present a framework to systematically analyze convolutional neural networks (CNNs) used in classification of cars in autonomous vehicles.

Autonomous Driving Classification +1

Logic-based Clustering and Learning for Time-Series Data

no code implementations22 Dec 2016 Marcell Vazquez-Chanlatte, Jyotirmoy V. Deshmukh, Xiaoqing Jin, Sanjit A. Seshia

To effectively analyze and design cyberphysical systems (CPS), designers today have to combat the data deluge problem, i. e., the burden of processing intractably large amounts of data produced by complex models and experiments.

Clustering General Classification +2

Towards Verified Artificial Intelligence

no code implementations27 Jun 2016 Sanjit A. Seshia, Dorsa Sadigh, S. Shankar Sastry

Verified artificial intelligence (AI) is the goal of designing AI-based systems that that have strong, ideally provable, assurances of correctness with respect to mathematically-specified requirements.

Constrained Sampling and Counting: Universal Hashing Meets SAT Solving

no code implementations21 Dec 2015 Kuldeep S. Meel, Moshe Vardi, Supratik Chakraborty, Daniel J. Fremont, Sanjit A. Seshia, Dror Fried, Alexander Ivrii, Sharad Malik

Constrained sampling and counting are two fundamental problems in artificial intelligence with a diverse range of applications, spanning probabilistic reasoning and planning to constrained-random verification.

Are There Good Mistakes? A Theoretical Analysis of CEGIS

no code implementations21 Jul 2014 Susmit Jha, Sanjit A. Seshia

The history bounded counterexample used in any iteration of CEGIS is bounded by the examples used in previous iterations of inductive synthesis.

Distribution-Aware Sampling and Weighted Model Counting for SAT

no code implementations11 Apr 2014 Supratik Chakraborty, Daniel J. Fremont, Kuldeep S. Meel, Sanjit A. Seshia, Moshe Y. Vardi

We present a novel approach that works with a black-box oracle for weights of assignments and requires only an {\NP}-oracle (in practice, a SAT-solver) to solve both the counting and sampling problems.

Computational Efficiency

Cannot find the paper you are looking for? You can Submit a new open access paper.