no code implementations • 7 Mar 2024 • Subbarao Kambhampati
While humans sometimes do show the capability of correcting their own erroneous guesses with self-critiquing, there seems to be no basis for that assumption in the case of LLMs.
no code implementations • 12 Feb 2024 • Kaya Stechly, Karthik Valmeekam, Subbarao Kambhampati
While the initial optimism that reasoning might emerge automatically with scale has been tempered thanks to a slew of counterexamples--ranging from multiplication to simple planning--there persists a wide spread belief that LLMs can self-critique and improve their own solutions in an iterative fashion.
no code implementations • 6 Feb 2024 • Lin Guan, Yifan Zhou, Denis Liu, Yantian Zha, Heni Ben Amor, Subbarao Kambhampati
Large-scale generative models are shown to be useful for sampling meaningful candidate solutions, yet they often overlook task constraints and user preferences.
no code implementations • 2 Feb 2024 • Subbarao Kambhampati, Karthik Valmeekam, Lin Guan, Kaya Stechly, Mudit Verma, Siddhant Bhambri, Lucas Saldyt, Anil Murthy
On the other side are perhaps over-pessimistic claims that all that LLMs are good for in planning/reasoning tasks are as mere translators of the problem specification from one syntactic format to another, and ship the problem off to external symbolic solvers.
no code implementations • 10 Jan 2024 • Mudit Verma, Siddhant Bhambri, Subbarao Kambhampati
In this work, we explore the task of Perceived Behavior Recognition, where a robot employs a Large Language Model (LLM) to assess the robot's generated behavior in a manner similar to human observer.
no code implementations • 21 Dec 2023 • Siddhant Bhambri, Mudit Verma, Anil Murthy, Subbarao Kambhampati
We introduce the notion of Human-Flexibility, i. e. whether the human partner is amenable to multiple team strategies, with a special case being Specified Orchestration where the human has a single team policy in mind (most constrained case).
no code implementations • 19 Oct 2023 • Kaya Stechly, Matthew Marquez, Subbarao Kambhampati
The study seems to indicate that (i) LLMs are bad at solving graph coloring instances (ii) they are no better at verifying a solution--and thus are not effective in iterative modes with LLMs critiquing LLM-generated solutions (iii) the correctness and content of the criticisms--whether by LLMs or external solvers--seems largely irrelevant to the performance of iterative prompting.
no code implementations • 12 Oct 2023 • Karthik Valmeekam, Matthew Marquez, Subbarao Kambhampati
We evaluate a planning system that employs LLMs for both plan generation and verification.
no code implementations • 26 May 2023 • Daman Arora, Subbarao Kambhampati
By randomly sampling actions from the same dataset, we generate examples of invalid actions which are then used to train a verifier which can check for action applicability.
2 code implementations • 25 May 2023 • Karthik Valmeekam, Matthew Marquez, Sarath Sreedharan, Subbarao Kambhampati
We aim to evaluate (1) the effectiveness of LLMs in generating plans autonomously in commonsense planning tasks and (2) the potential of LLMs in LLM-Modulo settings where they act as a source of heuristic guidance for external planners and verifiers.
no code implementations • 28 Feb 2023 • Tung Thai, Ming Shen, Mayank Garg, Ayush Kalani, Nakul Vaidya, Utkarsh Soni, Mudit Verma, Sriram Gopalakrishnan, Neeraj Varshney, Chitta Baral, Subbarao Kambhampati, Jivko Sinapov, Matthias Scheutz
Learning to detect, characterize and accommodate novelties is a challenge that agents operating in open-world domains need to address to be able to guarantee satisfactory task performance.
no code implementations • 17 Feb 2023 • Mudit Verma, Siddhant Bhambri, Subbarao Kambhampati
Preference Based Reinforcement Learning has shown much promise for utilizing human binary feedback on queried trajectory pairs to recover the underlying reward model of the Human in the Loop (HiL).
no code implementations • 17 Feb 2023 • Mudit Verma, Subbarao Kambhampati
We propose a data-driven reward initialization method that does not add any additional cost to the human in the loop and negligible cost to the PbRL agent and show that doing so ensures that the predicted rewards of the initialized reward model are uniform in the state space and this reduces the variability in the performance of the method across multiple runs and is shown to improve the overall performance compared to other initialization methods.
no code implementations • 17 Feb 2023 • Mudit Verma, Subbarao Kambhampati
Reinforcement Learning has suffered from poor reward specification, and issues for reward hacking even in simple enough domains.
no code implementations • 13 Feb 2023 • Karthik Valmeekam, Sarath Sreedharan, Matthew Marquez, Alberto Olmo, Subbarao Kambhampati
On this benchmark, we evaluate LLMs in three modes: autonomous, heuristic and human-in-the-loop.
no code implementations • 29 Jan 2023 • Zahra Zahedi, Sarath Sreedharan, Subbarao Kambhampati
Handling trust is one of the core requirements for facilitating effective interaction between the human and the AI agent.
no code implementations • 28 Oct 2022 • Lin Guan, Karthik Valmeekam, Subbarao Kambhampati
We propose two practical methods that can learn to model any kind of behavioral attributes from ordered behavior clips.
no code implementations • 27 Oct 2022 • Utkarsh Soni, Nupur Thakur, Sarath Sreedharan, Lin Guan, Mudit Verma, Matthew Marquez, Subbarao Kambhampati
If the relevant concept is not in the shared vocabulary, then it is learned.
no code implementations • 7 Oct 2022 • Mudit Verma, Ayush Kharkwal, Subbarao Kambhampati
Through our experiments, we show that our method can provide an interpretable means of solving the Advice-Conformance Verification problem by conveying whether or not the agent is using the human's advice.
2 code implementations • NeurIPS 2023 • Karthik Valmeekam, Matthew Marquez, Alberto Olmo, Sarath Sreedharan, Subbarao Kambhampati
PlanBench provides sufficient diversity in both the task domains and the specific planning capabilities.
no code implementations • 18 Feb 2022 • Zahra Zahedi, Sarath Sreedharan, Subbarao Kambhampati
Through this paper, we will see how this new framework allows us to capture the various works done in the space of human-AI interaction and identify the fundamental behavioral patterns supported by these works.
1 code implementation • 6 Feb 2022 • Lin Guan, Sarath Sreedharan, Subbarao Kambhampati
At the low level, we learn a set of diverse policies for each possible task subgoal identified by the landmark, which are then stitched together.
no code implementations • 19 Oct 2021 • Kebing Jin, Hankz Hankui Zhuo, Zhanhao Xiao, Hai Wan, Subbarao Kambhampati
In this paper, we propose a novel algorithm framework to solve numeric planning problems mixed with logical relations and numeric changes based on gradient descent.
1 code implementation • 11 Oct 2021 • Yantian Zha, Lin Guan, Subbarao Kambhampati
Our main contribution is to propose the Self-Explanation for RL from Demonstrations (SERLfD) framework, which can overcome the limitations of traditional RLfD works.
no code implementations • 21 Sep 2021 • Subbarao Kambhampati, Sarath Sreedharan, Mudit Verma, Yantian Zha, Lin Guan
The jury is still out on whether AI systems will need to use symbols in their internal reasoning to achieve general intelligence capabilities.
1 code implementation • 15 Sep 2021 • Sriram Gopalakrishnan, Mudit Verma, Subbarao Kambhampati
We present a framework to model the human agent's behavior with respect to state uncertainty, and can be used to compute MDP policies that accounts for these problems.
no code implementations • 9 Jul 2021 • Sriram Gopalakrishnan, Utkarsh Soni, Tung Thai, Panagiotis Lymperopoulos, Matthias Scheutz, Subbarao Kambhampati
The game of monopoly is an adversarial multi-agent domain where there is no fixed goal other than to be the last player solvent, There are useful subgoals like monopolizing sets of properties, and developing them.
no code implementations • 23 Jun 2021 • Utkarsh Soni, Sarath Sreedharan, Subbarao Kambhampati
The former is achieved by a data-driven clustering approach while for the latter, we compile our explanation generation problem into a POMDP.
1 code implementation • 14 Jun 2021 • Alberto Olmo, Sarath Sreedharan, Subbarao Kambhampati
Operations in many essential industries including finance and banking are often characterized by the need to perform repetitive sequential tasks.
no code implementations • 3 May 2021 • Zahra Zahedi, Mudit Verma, Sarath Sreedharan, Subbarao Kambhampati
The problem of trust management is particularly challenging in mixed human-robot teams where the human and the robot may have different models about the task at hand and thus may have different expectations regarding the current course of action, thereby forcing the robot to focus on the costly explicable behavior.
no code implementations • 2 May 2021 • Anagha Kulkarni, Siddharth Srivastava, Subbarao Kambhampati
This paper addresses the problem of synthesizing the behavior of an AI agent that provides proactive task assistance to a human in settings like factory floors where they may coexist in a common environment.
no code implementations • 21 Apr 2021 • Sarath Sreedharan, Anagha Kulkarni, David E. Smith, Subbarao Kambhampati
Existing approaches for generating human-aware agent behaviors have considered different measures of interpretability in isolation.
no code implementations • 18 Mar 2021 • Zahra Zahedi, Subbarao Kambhampati
In this paper, we aim at providing a comprehensive outline of the different threads of work in human-AI collaboration.
no code implementations • 1 Jan 2021 • Sailik Sengupta, Subbarao Kambhampati
We argue that existing models are inadequate in sequential settings when there is incomplete information about rational adversary and yield sub-optimal movement strategies.
no code implementations • 24 Nov 2020 • Sachin Grover, David Smith, Subbarao Kambhampati
We show how to generate questions to refine the robot's understanding of the teammate's model.
no code implementations • 22 Nov 2020 • Sarath Sreedharan, Anagha Kulkarni, Tathagata Chakraborti, David E. Smith, Subbarao Kambhampati
Existing approaches for the design of interpretable agent behavior consider different measures of interpretability in isolation.
no code implementations • 19 Nov 2020 • Karthik Valmeekam, Sarath Sreedharan, Sailik Sengupta, Subbarao Kambhampati
Decision support systems seek to enable informed decision-making.
no code implementations • 28 Oct 2020 • Sriram Gopalakrishnan, Subbarao Kambhampati
In situations where humans and robots are moving in the same space whilst performing their own tasks, predictable paths taken by mobile robots can not only make the environment feel safer, but humans can also help with the navigation in the space by avoiding path conflicts or not blocking the way.
1 code implementation • 8 Oct 2020 • Sailik Sengupta, Kaustav Basu, Arunabha Sen, Subbarao Kambhampati
In this paper, we draw inspiration from work in Moving Target Defense (MTD) and consider a dynamic monitoring strategy that makes it difficult for an attacker to prevent unique identification of behavioral signals that indicate the status of HVTs.
Computer Science and Game Theory
no code implementations • 20 Jul 2020 • Sailik Sengupta, Subbarao Kambhampati
We argue that existing models are inadequate in sequential settings when there is incomplete information about a rational adversary and yield sub-optimal movement strategies.
no code implementations • 2 Jul 2020 • Anagha Kulkarni, Sarath Sreedharan, Sarah Keren, Tathagata Chakraborti, David Smith, Subbarao Kambhampati
Given structured environments (like warehouses and restaurants), it may be possible to design the environment so as to boost the interpretability of the robot's behavior or to shape the human's expectations of the robot's behavior.
no code implementations • 26 Jun 2020 • Alberto Olmo, Sailik Sengupta, Subbarao Kambhampati
classifying the image of a dog to an airplane) can perplex humans and result in the loss of human trust in the system.
1 code implementation • NeurIPS 2021 • Lin Guan, Mudit Verma, Sihang Guo, Ruohan Zhang, Subbarao Kambhampati
We focus on the task of learning from feedback, in which the human trainer not only gives binary evaluative "good" or "bad" feedback for queried state-action pairs, but also provides a visual explanation by annotating relevant features in images.
no code implementations • 26 Feb 2020 • Tathagata Chakraborti, Sarath Sreedharan, Subbarao Kambhampati
In this paper, we provide a comprehensive outline of the different threads of work in Explainable AI Planning (XAIP) that has emerged as a focus area in the last couple of years and contrast that with earlier efforts in the field in terms of techniques, target users, and delivery mechanisms.
no code implementations • 5 Feb 2020 • Zahra Zahedi, Sailik Sengupta, Subbarao Kambhampati
Task allocation is an important problem in multi-agent systems.
no code implementations • ICLR 2022 • Sarath Sreedharan, Utkarsh Soni, Mudit Verma, Siddharth Srivastava, Subbarao Kambhampati
As increasingly complex AI systems are introduced into our daily lives, it becomes important for such systems to be capable of explaining the rationale for their decisions and allowing users to contest these decisions.
no code implementations • 26 Jan 2020 • Niharika Jain, Alberto Olmo, Sailik Sengupta, Lydia Manikonda, Subbarao Kambhampati
In this paper, we show that popular Generative Adversarial Networks (GANs) exacerbate biases along the axes of gender and skin tone when given a skewed distribution of face-shots.
no code implementations • 15 Oct 2019 • Subbarao Kambhampati
From its inception, AI has had a rather ambivalent relationship to humans---swinging between their augmentation and replacement.
no code implementations • 26 Aug 2019 • Hankz Hankui Zhuo, Jing Peng, Subbarao Kambhampati
Our approach takes as input a set of plan traces with disordered actions and noise and outputs action models that can best explain the plan traces.
no code implementations • 25 May 2019 • Anagha Kulkarni, Siddharth Srivastava, Subbarao Kambhampati
In order to be useful in the real world, AI agents need to plan and act in the presence of others, who may include adversarial and cooperative entities.
no code implementations • 19 Mar 2019 • Sarath Sreedharan, Siddharth Srivastava, David Smith, Subbarao Kambhampati
Explainable planning is widely accepted as a prerequisite for autonomous agents to successfully work with humans.
no code implementations • 18 Mar 2019 • Sarath Sreedharan, Tathagata Chakraborti, Christian Muise, Subbarao Kambhampati
In this work, we present a new planning formalism called Expectation-Aware planning for decision making with humans in the loop where the human's expectations about an agent may differ from the agent's own model.
no code implementations • 17 Mar 2019 • Sarath Sreedharan, Alberto Olmo, Aditya Prasad Mishra, Subbarao Kambhampati
One such approach has been the idea of {\em explanation as model-reconciliation}.
no code implementations • 1 Mar 2019 • Zahra Zahedi, Sailik Sengupta, Subbarao Kambhampati
Thus, we define the concept of a trust boundary over the mixed strategy space of the human and show that it helps to discover optimal monitoring strategies.
2 code implementations • 23 Dec 2018 • Ankur Chowdhary, Sailik Sengupta, Dijiang Huang, Subbarao Kambhampati
The processing and storage of critical data in large-scale cloud networks necessitate the need for scalable security solutions.
no code implementations • 2 Dec 2018 • Yantian Zha, Yikang Li, Tianshu Yu, Subbarao Kambhampati, Baoxin Li
We build an event recognition system, ER-PRN, which takes Pixel Dynamics Network as a subroutine, to recognize events based on observations augmented by plan-recognition-driven attention.
no code implementations • 24 Nov 2018 • Sriram Gopalakrishnan, Subbarao Kambhampati
TGE-viz allows users to visualize and criticize plans more intuitively for mixed-initiative planning.
no code implementations • 23 Nov 2018 • Tathagata Chakraborti, Anagha Kulkarni, Sarath Sreedharan, David E. Smith, Subbarao Kambhampati
There has been significant interest of late in generating behavior of agents that is interpretable to the human (observer) in the loop.
no code implementations • 9 Nov 2018 • Niharika Jain, Lydia Manikonda, Alberto Olmo Hernandez, Sailik Sengupta, Subbarao Kambhampati
The use of synthetic data generated by Generative Adversarial Networks (GANs) has become quite a popular method to do data augmentation for many applications.
no code implementations • 7 Mar 2018 • Wenfeng Feng, Hankz Hankui Zhuo, Subbarao Kambhampati
Extracting action sequences from natural language texts is challenging, as it requires commonsense inferences based on world knowledge.
no code implementations • 4 Mar 2018 • Hankz Hankui Zhuo, Yantian Zha, Subbarao Kambhampati
Specifically, we propose two approaches, DUP and RNNPlanner, to discover target plans based on vector representations of actions.
no code implementations • 19 Feb 2018 • Sarath Sreedharan, Siddharth Srivastava, Subbarao Kambhampati
There is a growing interest within the AI research community to develop autonomous systems capable of explaining their behavior to users.
no code implementations • 16 Feb 2018 • Anagha Kulkarni, Siddharth Srivastava, Subbarao Kambhampati
By slightly varying our framework, we present an approach for goal legibility in cooperative settings which produces plans that achieve a goal while being consistent with at most j goals from a set of confounding goals.
no code implementations • 3 Feb 2018 • Tathagata Chakraborti, Sarath Sreedharan, Sachin Grover, Subbarao Kambhampati
Recent work in explanation generation for decision making agents has looked at how unexplained behavior of autonomous systems can be understood in terms of differences in the model of the system and the human's understanding of the same, and how the explanation process as a result of this mismatch can be then seen as a process of reconciliation of these models.
no code implementations • 30 Jan 2018 • Tathagata Chakraborti, Subbarao Kambhampati
Effective collaboration between humans and AI-based systems requires effective modeling of the human in the loop, both in terms of the mental state as well as the physical capabilities of the latter.
no code implementations • 5 Dec 2017 • Yantian Zha, Yikang Li, Sriram Gopalakrishnan, Baoxin Li, Subbarao Kambhampati
The first involves resampling the distribution sequences to single action sequences, from which we could learn an action affinity model based on learned action (word) embeddings for plan recognition.
no code implementations • 25 Sep 2017 • Lydia Manikonda, Subbarao Kambhampati
Specifically, this paper performs a comparative analysis on the understanding of users belonging to two categories -- general AI-Tweeters (AIT) and expert AI-Tweeters (EAIT) who share posts about AI on Twitter.
no code implementations • 1 Aug 2017 • Tathagata Chakraborti, Sarath Sreedharan, Subbarao Kambhampati
In this paper, we bring these two concepts together and show how a planner can account for both these needs and achieve a trade-off during the plan generation process itself by means of a model-space search method MEGA.
no code implementations • 15 Jul 2017 • Tathagata Chakraborti, Subbarao Kambhampati, Matthias Scheutz, Yu Zhang
Among the many anticipated roles for robots in the future is that of being a human teammate.
1 code implementation • 19 May 2017 • Sailik Sengupta, Tathagata Chakraborti, Subbarao Kambhampati
Present attack methods can make state-of-the-art classification systems based on deep neural networks misclassify every adversarially modified test example.
no code implementations • 27 Apr 2017 • Lydia Manikonda, Cameron Dudley, Subbarao Kambhampati
Specifically, this paper performs a comparative analysis on the understanding of users from two categories -- general AI-Tweeters (AIT) and the expert AI-Tweeters (EAIT) who share posts about AI on Twitter.
no code implementations • 28 Jan 2017 • Tathagata Chakraborti, Sarath Sreedharan, Yu Zhang, Subbarao Kambhampati
When AI systems interact with humans in the loop, they are often called on to provide explanations for their plans and behavior.
no code implementations • 16 Nov 2016 • Anagha Kulkarni, Yantian Zha, Tathagata Chakraborti, Satya Gautam Vadlamudi, Yu Zhang, Subbarao Kambhampati
In order to have effective human-AI collaboration, it is necessary to address how the AI agent's behavior is being perceived by the humans-in-the-loop.
no code implementations • 27 Sep 2016 • Tathagata Chakraborti, Kartik Talamadupula, Kshitij P. Fadnis, Murray Campbell, Subbarao Kambhampati
In this paper, we present UbuntuWorld 1. 0 LTS - a platform for developing automated technical support agents in the Ubuntu operating system.
no code implementations • 24 Jun 2016 • Satya Gautam Vadlamudi, Tathagata Chakraborti, Yu Zhang, Subbarao Kambhampati
Proactive decision support (PDS) helps in improving the decision making experience of human decision makers in human-in-the-loop planning environments.
no code implementations • 25 May 2016 • Tathagata Chakraborti, Sarath Sreedharan, Sailik Sengupta, T. K. Satish Kumar, Subbarao Kambhampati
In this paper, we develop a computationally simpler version of the operator count heuristic for a particular class of domains.
1 code implementation • 23 Feb 2016 • Sailik Sengupta, Satya Gautam Vadlamudi, Subbarao Kambhampati, Marthony Taguinod, Adam Doupé, Ziming Zhao, Gail-Joon Ahn
We also address the issue of prioritizing vulnerabilities that when fixed, improves the security of the MTD system.
no code implementations • 25 Nov 2015 • Yu Zhang, Sarath Sreedharan, Anagha Kulkarni, Tathagata Chakraborti, Hankz Hankui Zhuo, Subbarao Kambhampati
Hence, for such agents to be helpful, one important requirement is for them to synthesize plans that can be easily understood by humans.
no code implementations • 18 Nov 2015 • Xin Tian, Hankz Hankui Zhuo, Subbarao Kambhampati
Previous approaches either discover plans by maximally "matching" observed actions to plan libraries, assuming target plans are from plan libraries, or infer plans by executing domain models to best explain the observed actions, assuming complete domain models are available.
no code implementations • 9 Dec 2014 • Vignesh Narayanan, Yu Zhang, Nathaniel Mendoza, Subbarao Kambhampati
While information asymmetry can be desirable sometimes, it may also lead to the robot choosing improper actions that negatively influence the teaming performance.
no code implementations • 4 Nov 2014 • Yu Zhang, Subbarao Kambhampati
Thus far, there are two common representations of agent models: MDP based and action based, which are both based on action modeling.
no code implementations • 1 Nov 2014 • William Cushing, J. Benton, Patrick Eyerich, Subbarao Kambhampati
Recently, several researchers have found that cost-based satisficing search with A* often runs into problems.
no code implementations • 12 May 2014 • Kartik Talamadupula, David E. Smith, Subbarao Kambhampati
An open question is whether these metrics are interchangeable; answering this requires a normalized comparison of the various replanning quality metrics.
no code implementations • 22 Apr 2014 • Yu Zhang, Subbarao Kambhampati
Then, by dividing the problems that require cooperation (referred to as RC problems) into two classes -- problems with heterogeneous and homogeneous agents, we aim to identify all the conditions that can cause RC in these two classes.
no code implementations • NeurIPS 2013 • Tuan A. Nguyen, Subbarao Kambhampati, Minh Do
In this paper, we first introduce annotations expressing the knowledge of the domain incompleteness and formalize the notion of plan robustness with respect to an incomplete domain model.
no code implementations • 29 Jul 2013 • Kartik Talamadupula, Subbarao Kambhampati
In this paper, we will argue that the automated oversight used in these systems can be viewed as a primitive automated planner, and that there are several opportunities for more sophisticated automated planning in effectively steering crowdsourced planning.
no code implementations • NeurIPS 2012 • Hankz H. Zhuo, Qiang Yang, Subbarao Kambhampati
Previous MAPR approaches required a library of team activity sequences (team plans) be given as input.