Kambhampati, Subbarao

The Emerging Landscape of Explainable AI Planning and Decision Making

arXiv.org Artificial Intelligence

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.

`Why not give this work to them?' Explaining AI-Moderated Task-Allocation Outcomes using Negotiation Trees

arXiv.org Artificial Intelligence

The problem of multi-agent task allocation arises in a variety of scenarios involving human teams. In many such settings, human teammates may act with selfish motives and try to minimize their cost metrics. In the absence of (1) complete knowledge about the reward of other agents and (2) the team's overall cost associated with a particular allocation outcome, distributed algorithms can only arrive at sub-optimal solutions within a reasonable amount of time. To address these challenges, we introduce the notion of an AI Task Allocator (AITA) that, with complete knowledge, comes up with fair allocations that strike a balance between the individual human costs and the team's performance cost. To ensure that AITA is explicable to the humans, we allow each human agent to question AITA's proposed allocation with counterfactual allocations. In response, we design AITA to provide a replay negotiation tree that acts as an explanation showing why the counterfactual allocation, with the correct costs, will eventually result in a sub-optimal allocation. This explanation also updates a human's incomplete knowledge about their teammate's and the team's actual costs. We then investigate whether humans are (1) able to understand the explanations provided and (2) convinced by it using human factor studies. Finally, we show the effect of various kinds of incompleteness on the length of explanations. We conclude that underestimation of other's costs often leads to the need for explanations and in turn, longer explanations on average.

Action-Model Based Multi-agent Plan Recognition

Neural Information Processing Systems

Multi-Agent Plan Recognition (MAPR) aims to recognize dynamic team structures and team behaviors from the observed team traces (activity sequences) of a set of intelligent agents. Previous MAPR approaches required a library of team activity sequences (team plans) be given as input. However, collecting a library of team plans to ensure adequate coverage is often difficult and costly. In this paper, we relax this constraint, so that team plans are not required to be provided beforehand. We assume instead that a set of action models are available.

Synthesizing Robust Plans under Incomplete Domain Models

Neural Information Processing Systems

Most current planners assume complete domain models and focus on generating correct plans. Unfortunately, domain modeling is a laborious and error-prone task, thus real world agents have to plan with incomplete domain models. While domain experts cannot guarantee completeness, often they are able to circumscribe the incompleteness of the model by providing annotations as to which parts of the domain model may be incomplete. In such cases, the goal should be to synthesize plans that are robust with respect to any known incompleteness of the domain. 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.

Bridging the Gap: Providing Post-Hoc Symbolic Explanations for Sequential Decision-Making Problems with Black Box Simulators

arXiv.org Artificial Intelligence

As more and more complex AI systems are introduced into our day-to-day lives, it becomes important that everyday users can work and interact with such systems with relative ease. Orchestrating such interactions require the system to be capable of providing explanations and rationale for its decisions and be able to field queries about alternative decisions. A significant hurdle to allowing for such explanatory dialogue could be the mismatch between the complex representations that the systems use to reason about the task and the terms in which the user may be viewing the task. This paper introduces methods that can be leveraged to provide contrastive explanations in terms of user-specified concepts for deterministic sequential decision-making settings where the system dynamics may be best represented in terms of black box simulators. We do this by assuming that system dynamics can at least be partly captured in terms of symbolic planning models, and we provide explanations in terms of these models. We implement this method using a simulator for a popular Atari game (Montezuma's Revenge) and perform user studies to verify whether people would find explanations generated in this form useful.

Imperfect ImaGANation: Implications of GANs Exacerbating Biases on Facial Data Augmentation and Snapchat Selfie Lenses

arXiv.org Machine Learning

Recently, the use of synthetic data generated by GANs has become a popular method to do data augmentation for many applications. While practitioners celebrate this as an economical way to obtain synthetic data for training data-hungry machine learning models, it is not clear that they recognize the perils of such an augmentation technique when applied to an already-biased dataset. Although one expects GANs to replicate the distribution of the original data, in real-world settings with limited data and finite network capacity, GANs suffer from mode collapse. Especially when this data is coming from online social media platforms or the web which are never balanced. In this paper, we show that in settings where data exhibits bias along some axes (eg. More often than not, this bias is unavoidable; we empirically demonstrate that given input of a dataset of headshots of engineering faculty collected from 47 online university directory webpages in the United States is biased toward white males, a state-of-the-art (unconditional variant of) GAN "imagines" faces of synthetic engineering professors that have masculine facial features and white skin color (inferred using human studies and a state-of- the-art gender recognition system). We also conduct a preliminary case study to highlight how Snapchat's explosively popular "female" filter (widely accepted to use a conditional variant of GAN), ends up consistently lightening the skin tones in women of color when trying to make face images appear more feminine. Our study is meant to serve as a cautionary tale for the lay practitioners who may unknowingly increase the bias in their training data by using GAN-based augmentation techniques with web data and to showcase the dangers of using biased datasets for facial applications. Introduction Breakthroughs in deep learning for image recognition have heralded significant progress in the field of computer vision, but one of the greatest limitations of the technology still remains: classifiers require massive amounts of training data to recognize meaningful patterns. As practitioners struggle to coax classifiers into generalizing to the underlying real-world distribution instead of overfitting to the sparse data, several data augmentation techniques have emerged as a useful form of regularization to increase training sample size and thereby, accuracy during test-time. The simplest of these augmentation techniques perform affine transformations on existing samples in the data, such as rotation, zooming, translation etc. (O'Gorman and Kasturi 1995; Bloice, Stocker, and Holzinger 2017). Ideally, the transformed samples should be representative of the same real-world distribution p dataas the original train and test sets.

Challenges of Human-Aware AI Systems

arXiv.org Artificial Intelligence

From its inception, AI has had a rather ambivalent relationship to 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. To do this effectively, AI systems must pay more attention to aspects of intelligence that helped humans work with each other---including social intelligence. I will discuss the research challenges in designing such human-aware AI systems, including modeling the mental states of humans in the loop, recognizing their desires and intentions, providing proactive support, exhibiting explicable behavior, giving cogent explanations on demand, and engendering trust. I will survey the progress made so far on these challenges, and highlight some promising directions. I will also touch on the additional ethical quandaries that such systems pose. I will end by arguing that the quest for human-aware AI systems broadens the scope of AI enterprise, necessitates and facilitates true inter-disciplinary collaborations, and can go a long way towards increasing public acceptance of AI technologies.

Learning Action Models from Disordered and Noisy Plan Traces

arXiv.org Artificial Intelligence

There is increasing awareness in the planning community that the burden of specifying complete domain models is too high, which impedes the applicability of planning technology in many real-world domains. Although there have many learning systems that help automatically learning domain models, most existing work assumes that the input traces are completely correct. A more realistic situation is that the plan traces are disordered and noisy, such as plan traces described by natural language. In this paper we propose and evaluate an approach for doing this. 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. We use a MAX-SAT framework for learning, where the constraints are derived from the given plan traces. Unlike traditional action models learners, the states in plan traces can be partially observable and noisy as well as the actions in plan traces can be disordered and parallel. We demonstrate the effectiveness of our approach through a systematic empirical evaluation with both IPC domains and the real-world dataset extracted from natural language documents.

Balancing Goal Obfuscation and Goal Legibility in Settings with Cooperative and Adversarial Observers

arXiv.org Artificial Intelligence

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. In this paper, we consider the problem where an autonomous agent needs to act in a manner that clarifies its objectives to cooperative entities while preventing adversarial entities from inferring those objectives. We show that this problem is solvable when cooperative entities and adversarial entities use different types of sensors and/or prior knowledge. We develop two new solution approaches for computing such plans. One approach provides an optimal solution to the problem by using an IP solver to provide maximum obfuscation for adversarial entities while providing maximum legibility for cooperative entities in the environment, whereas the other approach provides a satisficing solution using heuristic-guided forward search to achieve preset levels of obfuscation and legibility for adversarial and cooperative entities respectively. We show the feasibility and utility of our algorithms through extensive empirical evaluation on problems derived from planning benchmarks.

Why Couldn't You do that? Explaining Unsolvability of Classical Planning Problems in the Presence of Plan Advice

arXiv.org Artificial Intelligence

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.