Sarath Sreedharan

Upasana Biswas, Durgesh Kalwar, Subbarao Kambhampati et al.

Mutual adaptation is a central challenge in human-AI teaming, as humans naturally adjust their strategies in response to an AI agent's behavior. Existing approaches attempt to approximate human behavior by diversifying training partners; however, these partners are typically static and fail to capture the adaptive nature of human teammates. When agents are trained jointly in standard multi-agent settings, they often converge to opaque coordination strategies that work only with their co-trained partners, leading to poor generalization. To model adaptive human behavior, we formulate human-AI teaming as an Interactive Partially Observable Markov Decision Process (I-POMDP). We propose NestRL, a nested training regime that learns the solution to a finite-level I-POMDP by training agents at each level against adaptive agents from the level below. This exposes agents to adaptive behavior while preventing emergence of opaque coordination strategies. We provide theoretical analysis showing that NestRL agents avoid convergence to partner-specific strategies, and validate this empirically in the Overcooked domain against state-of-the-art baselines. NestRL achieves higher task performance with both unseen adaptive agents and real human teammates, while exhibiting significantly greater adaptability over the course of interaction.

Karthik Valmeekam, Matthew Marquez, Alberto Olmo et al.

Generating plans of action, and reasoning about change have long been considered a core competence of intelligent agents. It is thus no surprise that evaluating the planning and reasoning capabilities of large language models (LLMs) has become a hot topic of research. Most claims about LLM planning capabilities are however based on common sense tasks-where it becomes hard to tell whether LLMs are planning or merely retrieving from their vast world knowledge. There is a strong need for systematic and extensible planning benchmarks with sufficient diversity to evaluate whether LLMs have innate planning capabilities. Motivated by this, we propose PlanBench, an extensible benchmark suite based on the kinds of domains used in the automated planning community, especially in the International Planning Competition, to test the capabilities of LLMs in planning or reasoning about actions and change. PlanBench provides sufficient diversity in both the task domains and the specific planning capabilities. Our studies also show that on many critical capabilities-including plan generation-LLM performance falls quite short, even with the SOTA models. PlanBench can thus function as a useful marker of progress of LLMs in planning and reasoning.

Karthik Valmeekam, Sarath Sreedharan, Matthew Marquez et al.

Intrigued by the claims of emergent reasoning capabilities in LLMs trained on general web corpora, in this paper, we set out to investigate their planning capabilities. We aim to evaluate (1) how good LLMs are by themselves in generating and validating simple plans in commonsense planning tasks (of the type that humans are generally quite good at) and (2) how good LLMs are in being a source of heuristic guidance for other agents--either AI planners or human planners--in their planning tasks. To investigate these questions in a systematic rather than anecdotal manner, we start by developing a benchmark suite based on the kinds of domains employed in the International Planning Competition. On this benchmark, we evaluate LLMs in three modes: autonomous, heuristic and human-in-the-loop. Our results show that LLM's ability to autonomously generate executable plans is quite meager, averaging only about 3% success rate. The heuristic and human-in-the-loop modes show slightly more promise. In addition to these results, we also make our benchmark and evaluation tools available to support investigations by research community.

Turgay Caglar, Sirine Belhaj, Tathagata Chakraborti et al.

This is the first work to look at the application of large language models (LLMs) for the purpose of model space edits in automated planning tasks. To set the stage for this union, we explore two different flavors of model space problems that have been studied in the AI planning literature and explore the effect of an LLM on those tasks. We empirically demonstrate how the performance of an LLM contrasts with combinatorial search (CS) -- an approach that has been traditionally used to solve model space tasks in planning, both with the LLM in the role of a standalone model space reasoner as well as in the role of a statistical signal in concert with the CS approach as part of a two-stage process. Our experiments show promising results suggesting further forays of LLMs into the exciting world of model space reasoning for planning tasks in the future.

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. Thus, any decision-making framework designed to work with humans must possess the ability to estimate and leverage human trust. In this paper, we propose a mental model based theory of trust that not only can be used to infer trust, thus providing an alternative to psychological or behavioral trust inference methods, but also can be used as a foundation for any trust-aware decision-making frameworks. First, we introduce what trust means according to our theory and then use the theory to define trust evolution, human reliance and decision making, and a formalization of the appropriate level of trust in the agent. Using human subject studies, we compare our theory against one of the most common trust scales (Muir scale) to evaluate 1) whether the observations from the human studies match our proposed theory and 2) what aspects of trust are more aligned with our proposed theory.

Utkarsh Soni, Nupur Thakur, Sarath Sreedharan et al.

There is a growing interest in developing automated agents that can work alongside humans. In addition to completing the assigned task, such an agent will undoubtedly be expected to behave in a manner that is preferred by the human. This requires the human to communicate their preferences to the agent. To achieve this, the current approaches either require the users to specify the reward function or the preference is interactively learned from queries that ask the user to compare behavior. The former approach can be challenging if the internal representation used by the agent is inscrutable to the human while the latter is unnecessarily cumbersome for the user if their preference can be specified more easily in symbolic terms. In this work, we propose PRESCA (PREference Specification through Concept Acquisition), a system that allows users to specify their preferences in terms of concepts that they understand. PRESCA maintains a set of such concepts in a shared vocabulary. If the relevant concept is not in the shared vocabulary, then it is learned. To make learning a new concept more feedback efficient, PRESCA leverages causal associations between the target concept and concepts that are already known. In addition, we use a novel data augmentation approach to further reduce required feedback. We evaluate PRESCA by using it on a Minecraft environment and show that it can effectively align the agent with the user's preference.

Malek Mechergui, Sarath Sreedharan

Value alignment problems arise in scenarios where the specified objectives of an AI agent don't match the true underlying objective of its users. The problem has been widely argued to be one of the central safety problems in AI. Unfortunately, most existing works in value alignment tend to focus on issues that are primarily related to the fact that reward functions are an unintuitive mechanism to specify objectives. However, the complexity of the objective specification mechanism is just one of many reasons why the user may have misspecified their objective. A foundational cause for misalignment that is being overlooked by these works is the inherent asymmetry in human expectations about the agent's behavior and the behavior generated by the agent for the specified objective. To address this lacuna, we propose a novel formulation for the value alignment problem, named goal alignment that focuses on a few central challenges related to value alignment. In doing so, we bridge the currently disparate research areas of value alignment and human-aware planning. Additionally, we propose a first-of-its-kind interactive algorithm that is capable of using information generated under incorrect beliefs about the agent, to determine the true underlying goal of the user.

Brittany Cates, Anagha Kulkarni, Sarath Sreedharan

In this paper, we propose a planning framework to generate a defense strategy against an attacker who is working in an environment where a defender can operate without the attacker's knowledge. The objective of the defender is to covertly guide the attacker to a trap state from which the attacker cannot achieve their goal. Further, the defender is constrained to achieve its goal within K number of steps, where K is calculated as a pessimistic lower bound within which the attacker is unlikely to suspect a threat in the environment. Such a defense strategy is highly useful in real world systems like honeypots or honeynets, where an unsuspecting attacker interacts with a simulated production system while assuming it is the actual production system. Typically, the interaction between an attacker and a defender is captured using game theoretic frameworks. Our problem formulation allows us to capture it as a much simpler infinite horizon discounted MDP, in which the optimal policy for the MDP gives the defender's strategy against the actions of the attacker. Through empirical evaluation, we show the merits of our problem formulation.

Malek Mechergui, Sarath Sreedharan

Detecting and handling misspecified objectives, such as reward functions, has been widely recognized as one of the central challenges within the domain of Artificial Intelligence (AI) safety research. However, even with the recognition of the importance of this problem, we are unaware of any works that attempt to provide a clear definition for what constitutes (a) misspecified objectives and (b) successfully resolving such misspecifications. In this work, we use the theory of mind, i.e., the human user's beliefs about the AI agent, as a basis to develop a formal explanatory framework called Expectation Alignment (EAL) to understand the objective misspecification and its causes. Our EAL framework not only acts as an explanatory framework for existing works but also provides us with concrete insights into the limitations of existing methods to handle reward misspecification and novel solution strategies. We use these insights to propose a new interactive algorithm that uses the specified reward to infer potential user expectations about the system behavior. We show how one can efficiently implement this algorithm by mapping the inference problem into linear programs. We evaluate our method on a set of standard Markov Decision Process (MDP) benchmarks.

Kelsey Sikes, Sarah Keren, Sarath Sreedharan

One of the most difficult challenges in creating successful human-AI collaborations is aligning a robot's behavior with a human user's expectations. When this fails to occur, a robot may misinterpret their specified goals, prompting it to perform actions with unanticipated, potentially dangerous side effects. To avoid this, we propose a new metric we call Goal State Divergence $\mathcal{(GSD)}$, which represents the difference between a robot's final goal state and the one a human user expected. In cases where $\mathcal{GSD}$ cannot be directly calculated, we show how it can be approximated using maximal and minimal bounds. We then input the $\mathcal{GSD}$ value into our novel human-robot goal alignment (HRGA) design problem, which identifies a minimal set of environment modifications that can prevent mismatches like this. To show the effectiveness of $\mathcal{GSD}$ for reducing differences between human-robot goal states, we empirically evaluate our approach on several standard benchmarks.

Sarath Sreedharan, Kelsey Sikes, Nathaniel Blanchard et al.

The role that highly curated knowledge, provided by domain experts, could play in creating effective tutoring systems is often overlooked within the AI for education community. In this paper, we highlight this topic by discussing two ways such highly curated expert knowledge could help in creating novel educational systems. First, we will look at how one could use explainable AI (XAI) techniques to automatically create lessons. Most existing XAI methods are primarily aimed at debugging AI systems. However, we will discuss how one could use expert specified rules about solving specific problems along with novel XAI techniques to automatically generate lessons that could be provided to learners. Secondly, we will see how an expert specified curriculum for learning a target concept can help develop adaptive tutoring systems, that can not only provide a better learning experience, but could also allow us to use more efficient algorithms to create these systems. Finally, we will highlight the importance of such methods using a case study of creating a tutoring system for pollinator identification, where such knowledge could easily be elicited from experts.

Silvia Tulli, Stylianos Loukas Vasileiou, Sarath Sreedharan

"Human-aware" has become a popular keyword used to describe a particular class of AI systems that are designed to work and interact with humans. While there exists a surprising level of consistency among the works that use the label human-aware, the term itself mostly remains poorly understood. In this work, we retroactively try to provide an account of what constitutes a human-aware AI system. We see that human-aware AI is a design oriented paradigm, one that focuses on the need for modeling the humans it may interact with. Additionally, we see that this paradigm offers us intuitive dimensions to understand and categorize the kinds of interactions these systems might have with humans. We show the pedagogical value of these dimensions by using them as a tool to understand and review the current landscape of work related to human-AI systems that purport some form of human modeling. To fit the scope of a workshop paper, we specifically narrowed our review to papers that deal with sequential decision-making and were published in a major AI conference in the last three years. Our analysis helps identify the space of potential research problems that are currently being overlooked. We perform additional analysis on the degree to which these works make explicit reference to results from social science and whether they actually perform user-studies to validate their systems. We also provide an accounting of the various AI methods used by these works.

Michael Katz, Harsha Kokel, Sarath Sreedharan · ibm-research

There is a broad consensus that the inability to form long-term plans is one of the key limitations of current foundational models and agents. However, the existing planning benchmarks remain woefully inadequate to truly measure their planning capabilities. Most existing benchmarks either focus on loosely defined tasks like travel planning or end up leveraging existing domains and problems from international planning competitions. While the former tasks are hard to formalize and verify, the latter were specifically designed to test and challenge the weaknesses of existing automated planners. To address these shortcomings, we propose a procedure for creating a planning benchmark centered around the game called Countdown, where a player is expected to form a target number from a list of input numbers through arithmetic operations. We discuss how this problem meets many of the desiderata associated with an ideal benchmark for planning capabilities evaluation. Specifically, the domain allows for an intuitive, natural language description for each problem instance, it is computationally challenging (NP-complete), and the instance space is rich enough that we do not have to worry about memorization. We perform an extensive theoretical analysis, establishing the computational complexity result and demonstrate the advantage of our instance generation procedure over public benchmarks. We evaluate a variety of existing LLM-assisted planning methods on instances generated using our procedure. Our results show that, unlike other domains like 24 Game (a special case of Countdown), our proposed dynamic benchmark remains extremely challenging for existing LLM-based approaches.

Michael Katz, Harsha Kokel, Christian Muise et al. · ibm-research

In over sixty years since its inception, the field of planning has made significant contributions to both the theory and practice of building planning software that can solve a never-before-seen planning problem. This was done through established practices of rigorous design and evaluation of planning systems. It is our position that this rigor should be applied to the current trend of work on planning with large language models. One way to do so is by correctly incorporating the insights, tools, and data from the automated planning community into the design and evaluation of LLM-based planners. The experience and expertise of the planning community are not just important from a historical perspective; the lessons learned could play a crucial role in accelerating the development of LLM-based planners. This position is particularly important in light of the abundance of recent works that replicate and propagate the same pitfalls that the planning community has encountered and learned from. We believe that avoiding such known pitfalls will contribute greatly to the progress in building LLM-based planners and to planning in general.

Rakesh Podder, Turgay Caglar, Shadaab Kawnain Bashir et al.

Graph-based frameworks are often used in network hardening to help a cyber defender understand how a network can be attacked and how the best defenses can be deployed. However, incorporating network connectivity parameters in the attack graph, reasoning about the attack graph when we do not have access to complete information, providing system administrator suggestions in an understandable format, and allowing them to do what-if analysis on various scenarios and attacker motives is still missing. We fill this gap by presenting SPEAR, a formal framework with tool support for security posture evaluation and analysis that keeps human-in-the-loop. SPEAR uses the causal formalism of AI planning to model vulnerabilities and configurations in a networked system. It automatically converts network configurations and vulnerability descriptions into planning models expressed in the Planning Domain Definition Language (PDDL). SPEAR identifies a set of diverse security hardening strategies that can be presented in a manner understandable to the domain expert. These allow the administrator to explore the network hardening solution space in a systematic fashion and help evaluate the impact and compare the different solutions.

Silvia Tulli, Stylianos Loukas Vasileiou, Mohamed Chetouani et al.

One of the significant challenges to generating value-aligned behavior is to not only account for the specified user objectives but also any implicit or unspecified user requirements. The existence of such implicit requirements could be particularly common in settings where the user's understanding of the task model may differ from the agent's estimate of the model. Under this scenario, the user may incorrectly expect some agent behavior to be inevitable or guaranteed. This paper addresses such expectation mismatch in the presence of differing models by capturing the possibility of unspecified user subgoal in the context of a task captured as a Markov Decision Process (MDP) and querying for it as required. Our method identifies bottleneck states and uses them as candidates for potential implicit subgoals. We then introduce a querying strategy that will generate the minimal number of queries required to identify a policy guaranteed to achieve the underlying goal. Our empirical evaluations demonstrate the effectiveness of our approach in inferring and achieving unstated goals across various tasks.

Sarath Sreedharan, Anagha Kulkarni, Subbarao Kambhampati

From its inception, AI has had a rather ambivalent relationship with humans -- swinging between their augmentation and replacement. Now, as AI technologies enter our everyday lives at an ever increasing pace, there is a greater need for AI systems to work synergistically with humans. One critical requirement for such synergistic human-AI interaction is that the AI systems be explainable to the humans in the loop. To do this effectively, AI agents need to go beyond planning with their own models of the world, and take into account the mental model of the human in the loop. Drawing from several years of research in our lab, we will discuss how the AI agent can use these mental models to either conform to human expectations, or change those expectations through explanatory communication. While the main focus of the book is on cooperative scenarios, we will point out how the same mental models can be used for obfuscation and deception. Although the book is primarily driven by our own research in these areas, in every chapter, we will provide ample connections to relevant research from other groups.

Karthik Valmeekam, Matthew Marquez, Sarath Sreedharan et al.

Intrigued by the claims of emergent reasoning capabilities in LLMs trained on general web corpora, in this paper, we set out to investigate their planning capabilities. 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. We conduct a systematic study by generating a suite of instances on domains similar to the ones employed in the International Planning Competition and evaluate LLMs in two distinct modes: autonomous and heuristic. Our findings reveal that LLMs' ability to generate executable plans autonomously is rather limited, with the best model (GPT-4) having an average success rate of ~12% across the domains. However, the results in the LLM-Modulo setting show more promise. In the LLM-Modulo setting, we demonstrate that LLM-generated plans can improve the search process for underlying sound planners and additionally show that external verifiers can help provide feedback on the generated plans and back-prompt the LLM for better plan generation.

Lin Guan, Karthik Valmeekam, Sarath Sreedharan et al.

There is a growing interest in applying pre-trained large language models (LLMs) to planning problems. However, methods that use LLMs directly as planners are currently impractical due to several factors, including limited correctness of plans, strong reliance on feedback from interactions with simulators or even the actual environment, and the inefficiency in utilizing human feedback. In this work, we introduce a novel alternative paradigm that constructs an explicit world (domain) model in planning domain definition language (PDDL) and then uses it to plan with sound domain-independent planners. To address the fact that LLMs may not generate a fully functional PDDL model initially, we employ LLMs as an interface between PDDL and sources of corrective feedback, such as PDDL validators and humans. For users who lack a background in PDDL, we show that LLMs can translate PDDL into natural language and effectively encode corrective feedback back to the underlying domain model. Our framework not only enjoys the correctness guarantee offered by the external planners but also reduces human involvement by allowing users to correct domain models at the beginning, rather than inspecting and correcting (through interactive prompting) every generated plan as in previous work. On two IPC domains and a Household domain that is more complicated than commonly used benchmarks such as ALFWorld, we demonstrate that GPT-4 can be leveraged to produce high-quality PDDL models for over 40 actions, and the corrected PDDL models are then used to successfully solve 48 challenging planning tasks. Resources, including the source code, are released at: https://guansuns.github.io/pages/llm-dm.

Zahra Zahedi, Sarath Sreedharan, Subbarao Kambhampati

There has been significant recent interest in developing AI agents capable of effectively interacting and teaming with humans. While each of these works try to tackle a problem quite central to the problem of human-AI interaction, they tend to rely on myopic formulations that obscure the possible inter-relatedness and complementarity of many of these works. The human-aware AI framework was a recent effort to provide a unified account for human-AI interaction by casting them in terms of their relationship to various mental models. Unfortunately, the current accounts of human-aware AI are insufficient to explain the landscape of the work doing in the space of human-AI interaction due to their focus on limited settings. In this paper, we aim to correct this shortcoming by introducing a significantly general version of human-aware AI interaction scheme, called generalized human-aware interaction (GHAI), that talks about (mental) models of six types. 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. We will also use this framework to identify potential gaps in the current literature and suggest future research directions to address these shortcomings.

Lin Guan, Sarath Sreedharan, Subbarao Kambhampati

Creating reinforcement learning (RL) agents that are capable of accepting and leveraging task-specific knowledge from humans has been long identified as a possible strategy for developing scalable approaches for solving long-horizon problems. While previous works have looked at the possibility of using symbolic models along with RL approaches, they tend to assume that the high-level action models are executable at low level and the fluents can exclusively characterize all desirable MDP states. Symbolic models of real world tasks are however often incomplete. To this end, we introduce Approximate Symbolic-Model Guided Reinforcement Learning, wherein we will formalize the relationship between the symbolic model and the underlying MDP that will allow us to characterize the incompleteness of the symbolic model. We will use these models to extract high-level landmarks that will be used to decompose the task. 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. We evaluate our system by testing on three different benchmark domains and show how even with incomplete symbolic model information, our approach is able to discover the task structure and efficiently guide the RL agent towards the goal.

Subbarao Kambhampati, Sarath Sreedharan, Mudit Verma et al.

Despite the surprising power of many modern AI systems that often learn their own representations, there is significant discontent about their inscrutability and the attendant problems in their ability to interact with humans. While alternatives such as neuro-symbolic approaches have been proposed, there is a lack of consensus on what they are about. There are often two independent motivations (i) symbols as a lingua franca for human-AI interaction and (ii) symbols as system-produced abstractions used by the AI system in its internal reasoning. The jury is still out on whether AI systems will need to use symbols in their internal reasoning to achieve general intelligence capabilities. Whatever the answer there is, the need for (human-understandable) symbols in human-AI interaction seems quite compelling. Symbols, like emotions, may well not be sine qua non for intelligence per se, but they will be crucial for AI systems to interact with us humans -- as we can neither turn off our emotions nor get by without our symbols. In particular, in many human-designed domains, humans would be interested in providing explicit (symbolic) knowledge and advice -- and expect machine explanations in kind. This alone requires AI systems to to maintain a symbolic interface for interaction with humans. In this blue sky paper, we argue this point of view, and discuss research directions that need to be pursued to allow for this type of human-AI interaction.

Utkarsh Soni, Sarath Sreedharan, Subbarao Kambhampati

There is a growing interest in designing autonomous agents that can work alongside humans. Such agents will undoubtedly be expected to explain their behavior and decisions. While generating explanations is an actively researched topic, most works tend to focus on methods that generate explanations that are one size fits all. As in the specifics of the user-model are completely ignored. The handful of works that look at tailoring their explanation to the user's background rely on having specific models of the users (either analytic models or learned labeling models). The goal of this work is thus to propose an end-to-end adaptive explanation generation system that begins by learning the different types of users that the agent could interact with. Then during the interaction with the target user, it is tasked with identifying the type on the fly and adjust its explanations accordingly. The former is achieved by a data-driven clustering approach while for the latter, we compile our explanation generation problem into a POMDP. We demonstrate the usefulness of our system on two domains using state-of-the-art POMDP solvers. We also report the results of a user study that investigates the benefits of providing personalized explanations in a human-robot interaction setting.

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. Despite their criticality to the business, workflows are rarely fully automated or even formally specified, though there may exist a number of natural language documents describing these procedures for the employees of the company. Plan extraction methods provide us with the possibility of extracting structure plans from such natural language descriptions of the plans/workflows, which could then be leveraged by an automated system. In this paper, we investigate the utility of generalized language models in performing such extractions directly from such texts. Such models have already been shown to be quite effective in multiple translation tasks, and our initial results seem to point to their effectiveness also in the context of plan extractions. Particularly, we show that GPT-3 is able to generate plan extraction results that are comparable to many of the current state of the art plan extraction methods.

Zahra Zahedi, Mudit Verma, Sarath Sreedharan et al.

Trust between team members is an essential requirement for any successful cooperation. Thus, engendering and maintaining the fellow team members' trust becomes a central responsibility for any member trying to not only successfully participate in the task but to ensure the team achieves its goals. 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 and forcing the robot to focus on the costly explicable behavior. We propose a computational model for capturing and modulating trust in such longitudinal human-robot interaction, where the human adopts a supervisory role. In our model, the robot integrates human's trust and their expectations from the robot into its planning process to build and maintain trust over the interaction horizon. By establishing the required level of trust, the robot can focus on maximizing the team goal by eschewing explicit explanatory or explicable behavior without worrying about the human supervisor monitoring and intervening to stop behaviors they may not necessarily understand. We model this reasoning about trust levels as a meta reasoning process over individual planning tasks. We additionally validate our model through a human subject experiment.

Sarath Sreedharan, Anagha Kulkarni, David E. Smith et al.

Existing approaches for generating human-aware agent behaviors have considered different measures of interpretability in isolation. Further, these measures have been studied under differing assumptions, thus precluding the possibility of designing a single framework that captures these measures under the same assumptions. In this paper, we present a unifying Bayesian framework that models a human observer's evolving beliefs about an agent and thereby define the problem of Generalized Human-Aware Planning. We will show that the definitions of interpretability measures like explicability, legibility and predictability from the prior literature fall out as special cases of our general framework. Through this framework, we also bring a previously ignored fact to light that the human-robot interactions are in effect open-world problems, particularly as a result of modeling the human's beliefs over the agent. Since the human may not only hold beliefs unknown to the agent but may also form new hypotheses about the agent when presented with novel or unexpected behaviors.

Sarath Sreedharan, Anagha Kulkarni, Tathagata Chakraborti et al.

Existing approaches for the design of interpretable agent behavior consider different measures of interpretability in isolation. In this paper we posit that, in the design and deployment of human-aware agents in the real world, notions of interpretability are just some among many considerations; and the techniques developed in isolation lack two key properties to be useful when considered together: they need to be able to 1) deal with their mutually competing properties; and 2) an open world where the human is not just there to interpret behavior in one specific form. To this end, we consider three well-known instances of interpretable behavior studied in existing literature -- namely, explicability, legibility, and predictability -- and propose a revised model where all these behaviors can be meaningfully modeled together. We will highlight interesting consequences of this unified model and motivate, through results of a user study, why this revision is necessary.

Sarath Sreedharan, Tathagata Chakraborti, Yara Rizk et al.

A new design of an AI assistant that has become increasingly popular is that of an "aggregated assistant" -- realized as an orchestrated composition of several individual skills or agents that can each perform atomic tasks. In this paper, we will talk about the role of planning in the automated composition of such assistants and explore how concepts in automated planning can help to establish transparency of the inner workings of the assistant to the end-user.

Karthik Valmeekam, Sarath Sreedharan, Sailik Sengupta et al.

Decision support systems seek to enable informed decision-making. In the recent years, automated planning techniques have been leveraged to empower such systems to better aid the human-in-the-loop. The central idea for such decision support systems is to augment the capabilities of the human-in-the-loop with automated planning techniques and enhance the quality of decision-making. In addition to providing planning support, effective decision support systems must be able to provide intuitive explanations based on specific user queries for proposed decisions to its end users. Using this as motivation, we present our decision support system RADAR-X that showcases the ability to engage the user in an interactive explanatory dialogue by first enabling them to specify an alternative to a proposed decision (which we refer to as foils), and then providing contrastive explanations to these user-specified foils which helps the user understand why a specific plan was chosen over the alternative (or foil). Furthermore, the system uses this dialogue to elicit the user's latent preferences and provides revised plan suggestions through three different interaction strategies.

Anagha Kulkarni, Sarath Sreedharan, Sarah Keren et al.

Designing robots capable of generating interpretable behavior is a prerequisite for achieving effective human-robot collaboration. This means that the robots need to be capable of generating behavior that aligns with human expectations and, when required, provide explanations to the humans in the loop. However, exhibiting such behavior in arbitrary environments could be quite expensive for robots, and in some cases, the robot may not even be able to exhibit the expected behavior. 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. In this paper, we investigate the opportunities and limitations of environment design as a tool to promote a type of interpretable behavior -- known in the literature as explicable behavior. We formulate a novel environment design framework that considers design over multiple tasks and over a time horizon. In addition, we explore the longitudinal aspect of explicable behavior and the trade-off that arises between the cost of design and the cost of generating explicable behavior over a time horizon.

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. We hope that the survey will provide guidance to new researchers in automated planning towards the role of explanations in the effective design of human-in-the-loop systems, as well as provide the established researcher with some perspective on the evolution of the exciting world of explainable planning.

Sarath Sreedharan, Utkarsh Soni, Mudit Verma et al.

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. A significant hurdle to allowing for such explanatory dialogue could be the vocabulary mismatch between the user and the AI system. This paper introduces methods for providing contrastive explanations in terms of user-specified concepts for sequential decision-making settings where the system's model of the task may be best represented as an inscrutable model. We do this by building partial symbolic models of a local approximation of the task that can be leveraged to answer the user queries. We test these methods on a popular Atari game (Montezuma's Revenge) and variants of Sokoban (a well-known planning benchmark) and report the results of user studies to evaluate whether people find explanations generated in this form useful.

Sarath Sreedharan, Siddharth Srivastava, David Smith et al.

Explainable planning is widely accepted as a prerequisite for autonomous agents to successfully work with humans. While there has been a lot of research on generating explanations of solutions to planning problems, explaining the absence of solutions remains an open and under-studied problem, even though such situations can be the hardest to understand or debug. In this paper, we show that hierarchical abstractions can be used to efficiently generate reasons for unsolvability of planning problems. In contrast to related work on computing certificates of unsolvability, we show that these methods can generate compact, human-understandable reasons for unsolvability. Empirical analysis and user studies show the validity of our methods as well as their computational efficacy on a number of benchmark planning domains.

Sarath Sreedharan, Tathagata Chakraborti, Christian Muise et al.

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. We show how this formulation allows agents to not only leverage existing strategies for handling model differences but can also exhibit novel behaviors that are generated through the combination of these different strategies. Our formulation also reveals a deep connection to existing approaches in epistemic planning. Specifically, we show how we can leverage classical planning compilations for epistemic planning to solve Expectation-Aware planning problems. To the best of our knowledge, the proposed formulation is the first complete solution to decision-making in the presence of diverging user expectations that is amenable to a classical planning compilation while successfully combining previous works on explanation and explicability. We empirically show how our approach provides a computational advantage over existing approximate approaches that unnecessarily try to search in the space of models while also failing to facilitate the full gamut of behaviors enabled by our framework.

Sarath Sreedharan, Alberto Olmo, Aditya Prasad Mishra et al.

Designing agents capable of explaining complex sequential decisions remain a significant open problem in automated decision-making. Recently, there has been a lot of interest in developing approaches for generating such explanations for various decision-making paradigms. One such approach has been the idea of {\em explanation as model-reconciliation}. The framework hypothesizes that one of the common reasons for the user's confusion could be the mismatch between the user's model of the task and the one used by the system to generate the decisions. While this is a general framework, most works that have been explicitly built on this explanatory philosophy have focused on settings where the model of user's knowledge is available in a declarative form. Our goal in this paper is to adapt the model reconciliation approach to the cases where such user models are no longer explicitly provided. We present a simple and easy to learn labeling model that can help an explainer decide what information could help achieve model reconciliation between the user and the agent.

Tathagata Chakraborti, Anagha Kulkarni, Sarath Sreedharan et al.

There has been significant interest of late in generating behavior of agents that is interpretable to the human (observer) in the loop. However, the work in this area has typically lacked coherence on the topic, with proposed solutions for "explicable", "legible", "predictable" and "transparent" planning with overlapping, and sometimes conflicting, semantics all aimed at some notion of understanding what intentions the observer will ascribe to an agent by observing its behavior. This is also true for the recent works on "security" and "privacy" of plans which are also trying to answer the same question, but from the opposite point of view -- i.e. when the agent is trying to hide instead of revealing its intentions. This paper attempts to provide a workable taxonomy of relevant concepts in this exciting and emerging field of inquiry.

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. One aspect of the explanation generation problem that has yet to receive much attention is the task of explaining plans to users whose level of expertise differ from that of the explainer. We propose an approach for addressing this problem by representing the user's model as an abstraction of the domain model that the planner uses. We present algorithms for generating minimal explanations in cases where this abstract human model is not known. We reduce the problem of generating explanation to a search over the space of abstract models and investigate possible greedy approximations for minimal explanations. We also empirically show that our approach can efficiently compute explanations for a variety of problems.

Tathagata Chakraborti, Sarath Sreedharan, Sachin Grover et al.

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. Existing algorithms in such settings, while having been built on contrastive, selective and social properties of explanations as studied extensively in the psychology literature, have not, to the best of our knowledge, been evaluated in settings with actual humans in the loop. As such, the applicability of such explanations to human-AI and human-robot interactions remains suspect. In this paper, we set out to evaluate these explanation generation algorithms in a series of studies in a mock search and rescue scenario with an internal semi-autonomous robot and an external human commander. We demonstrate to what extent the properties of these algorithms hold as they are evaluated by humans, and how the dynamics of trust between the human and the robot evolve during the process of these interactions.

Tathagata Chakraborti, Sarath Sreedharan, Subbarao Kambhampati

Human aware planning requires an agent to be aware of the intentions, capabilities and mental model of the human in the loop during its decision process. This can involve generating plans that are explicable to a human observer as well as the ability to provide explanations when such plans cannot be generated. This has led to the notion "multi-model planning" which aim to incorporate effects of human expectation in the deliberative process of a planner - either in the form of explicable task planning or explanations produced thereof. 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. This in effect provides a comprehensive perspective of what it means for a decision making agent to be "human-aware" by bringing together existing principles of planning under the umbrella of a single plan generation process. We situate our discussion specifically keeping in mind the recent work on explicable planning and explanation generation, and illustrate these concepts in modified versions of two well known planning domains, as well as a demonstration on a robot involved in a typical search and reconnaissance task with an external supervisor.

Tathagata Chakraborti, Sarath Sreedharan, Anagha Kulkarni et al.

Ambiguity and noise in natural language instructions create a significant barrier towards adopting autonomous systems into safety critical workflows involving humans and machines. In this paper, we propose to build on recent advances in electrophysiological monitoring methods and augmented reality technologies, to develop alternative modes of communication between humans and robots involved in large-scale proximal collaborative tasks. We will first introduce augmented reality techniques for projecting a robot's intentions to its human teammate, who can interact with these cues to engage in real-time collaborative plan execution with the robot. We will then look at how electroencephalographic (EEG) feedback can be used to monitor human response to both discrete events, as well as longer term affective states while execution of a plan. These signals can be used by a learning agent, a.k.a an affective robot, to modify its policy. We will present an end-to-end system capable of demonstrating these modalities of interaction. We hope that the proposed system will inspire research in augmenting human-robot interactions with alternative forms of communications in the interests of safety, productivity, and fluency of teaming, particularly in engineered settings such as the factory floor or the assembly line in the manufacturing industry where the use of such wearables can be enforced.

Tathagata Chakraborti, Sarath Sreedharan, Yu Zhang et al.

When AI systems interact with humans in the loop, they are often called on to provide explanations for their plans and behavior. Past work on plan explanations primarily involved the AI system explaining the correctness of its plan and the rationale for its decision in terms of its own model. Such soliloquy is wholly inadequate in most realistic scenarios where the humans have domain and task models that differ significantly from that used by the AI system. We posit that the explanations are best studied in light of these differing models. In particular, we show how explanation can be seen as a "model reconciliation problem" (MRP), where the AI system in effect suggests changes to the human's model, so as to make its plan be optimal with respect to that changed human model. We will study the properties of such explanations, present algorithms for automatically computing them, and evaluate the performance of the algorithms.

Tathagata Chakraborti, Sarath Sreedharan, Sailik Sengupta et al.

In this paper, we develop a computationally simpler version of the operator count heuristic for a particular class of domains. The contribution of this abstract is threefold, we (1) propose an efficient closed form approximation to the operator count heuristic using the Lagrangian dual; (2) leverage compressed sensing techniques to obtain an integer approximation for operator counts in polynomial time; and (3) discuss the relationship of the proposed formulation to existing heuristics and investigate properties of domains where such approaches appear to be useful.

Yu Zhang, Sarath Sreedharan, Anagha Kulkarni et al.

Intelligent robots and machines are becoming pervasive in human populated environments. A desirable capability of these agents is to respond to goal-oriented commands by autonomously constructing task plans. However, such autonomy can add significant cognitive load and potentially introduce safety risks to humans when agents behave unexpectedly. Hence, for such agents to be helpful, one important requirement is for them to synthesize plans that can be easily understood by humans. While there exists previous work that studied socially acceptable robots that interact with humans in "natural ways", and work that investigated legible motion planning, there lacks a general solution for high level task planning. To address this issue, we introduce the notions of plan {\it explicability} and {\it predictability}. To compute these measures, first, we postulate that humans understand agent plans by associating abstract tasks with agent actions, which can be considered as a labeling process. We learn the labeling scheme of humans for agent plans from training examples using conditional random fields (CRFs). Then, we use the learned model to label a new plan to compute its explicability and predictability. These measures can be used by agents to proactively choose or directly synthesize plans that are more explicable and predictable to humans. We provide evaluations on a synthetic domain and with human subjects using physical robots to show the effectiveness of our approach