Arjun Akula (Arjun Reddy Akula)

We present a new explainable AI (XAI) framework aimed at increasing justified human trust and reliance in the AI machine through explanations. We pose explanation as an iterative communication process, i.e. dialog, between the machine and human user. More concretely, the machine generates sequence of explanations in a dialog which takes into account three important aspects at each dialog turn: (a) human's intention (or curiosity); (b) human's understanding of the machine; and (c) machine's understanding of the human user. To do this, we use Theory of Mind (ToM) which helps us in explicitly modeling human's intention, machine's mind as inferred by the human as well as human's mind as inferred by the machine. In other words, these explicit mental representations in ToM are incorporated to learn an optimal explanation policy that takes into account human's perception and beliefs. Furthermore, we also show that ToM facilitates in quantitatively measuring justified human trust in the machine by comparing all the three mental representations. We applied our framework to three visual recognition tasks, namely, image classification, action recognition, and human body pose estimation. We argue that our ToM based explanations are practical and more natural for both expert and non-expert users to understand the internal workings of complex machine learning models. To the best of our knowledge, this is the first work to derive explanations using ToM. Extensive human study experiments verify our hypotheses, showing that the proposed explanations significantly outperform the state-of-the-art XAI methods in terms of all the standard quantitative and qualitative XAI evaluation metrics including human trust, reliance, and explanation satisfaction.

We present CoCoX (short for Conceptual and Counterfactual Explanations), a model for explaining decisions made by a deep convolutional neural network (CNN). In Cognitive Psychology, the factors (or semantic-level features) that humans zoom in on when they imagine an alternative to a model prediction are often referred to as Fault-Lines. Motivated by this, our CoCoX model explains decisions made by a CNN using fault-lines. Specifically, given an input image I for which a CNN classification model M predicts class c_pred, our fault-line based explanation identifies the minimal semantic-level features (e.g., stripes on zebra, pointed ears of dog), referred to as explainable concepts, that need to be added to or deleted from I in order to alter the classification category of I by M to another specified class c_alt. We argue that, due to the conceptual and counterfactual nature of fault-lines, our CoCoX explanations are practical and more natural for both expert and non-expert users to understand the internal workings of complex deep learning models. Extensive quantitative and qualitative experiments verify our hypotheses, showing that CoCoX significantly outperform the state-of-the-art explainable AI models.

Visual referring expression recognition is a challenging task that requires natural language understanding in the context of an image. We critically examine RefCOCOg, a standard benchmark for this task, using a human study and show that 83.7% of test instances do not require reasoning on linguistic structure. To measure the true progress of existing models, we split the test set into two sets, one which requires reasoning on linguistic structure and the other which doesn't. Additionally, we create an adversarial dataset Ref-Adv that tests the model's ability to generalize to unseen distribution of target referring objects. Using these datasets, we empirically show that existing methods fail to exploit linguistic structure and are 12% to 23% lower in performance than the established progress for this task. We also propose two methods, one based on negative sampling and the other based on multi-task learning, to increase the robustness of ViLBERT, the current state-of-the-art model for this task.

Text-level discourse parsing aims to unmask how two segments (or sentences) in the text are related to each other. We propose the task of Visual Discourse Parsing, which requires understanding discourse relations among scenes in a video. Here we use the term scene to refer to a subset of video frames that can better summarize the video. In order to collect a dataset for learning discourse cues from videos, one needs to manually identify the scenes from a large pool of video frames and then annotate the discourse relations between them. This is clearly a time consuming, expensive and tedious task. In this work, we propose an approach to identify discourse cues from the videos without the need to explicitly identify and annotate the scenes. We also present a novel dataset containing 310 videos and the corresponding discourse cues to evaluate our approach. We believe that many of the multi-discipline Artificial Intelligence problems such as Visual Dialog and Visual Storytelling would greatly benefit from the use of visual discourse cues.

This paper presents an explainable AI (XAI) system that provides explanations for its predictions. The system consists of two key components -- namely, the prediction And-Or graph (AOG) model for recognizing and localizing concepts of interest in input data, and the XAI model for providing explanations to the user about the AOG's predictions. In this work, we focus on the XAI model specified to interact with the user in natural language, whereas the AOG's predictions are considered given and represented by the corresponding parse graphs (pg's) of the AOG. Our XAI model takes pg's as input and provides answers to the user's questions using the following types of reasoning: direct evidence (e.g., detection scores), part-based inference (e.g., detected parts provide evidence for the concept asked), and other evidences from spatio-temporal context (e.g., constraints from the spatio-temporal surround). We identify several correlations between user's questions and the XAI answers using Youtube Action dataset.

Many large IT service providers, having realized the transformational impact of cognitive technology, are experimenting with cognitive service agents like Amelia and IBM Watson for IT problem resolution. We build a system that extracts knowledge about different classes of problems arising in the IT infrastructure, mine problem linkages to recent system changes, and identify the resolution activities to mitigate problems. The system, at its core, uses data mining, machine learning, and natural language parsing techniques. By using extracted knowledge, one can (i) understand the kind of problems and the root causes affecting the IT infrastructure, (ii) proactively remediate the causes so that they no longer result in problems, and (iii) estimate the scope for automation for service management

Attribute information in a natural language query is one of the key features for converting a natural language query into a Structured Query Language (SQL) in Natural Language Interface to Database systems. In this paper, we explore the task of classifying the attributes present in a natural language query into different SQL clauses in a SQL query. In particular, we investigate the effectiveness of various features and Conditional Random Fields for this task. Our system uses a statistical classifier trained on manually prepared data. We report our results on three different domains and also show how our system can be used for generating a complete SQL query.

Service interactions account for major source of revenue and employment in many modern economies, and yet the service operations management process remains extremely complex. Ticket is the fundamental management entity in this process and resolution of tickets remains largely human intensive. A large portion of these human executed resolution tasks are repetitive in nature and can be automated. Ticket description analytics can be used to automatically identify the true category of the problem. This when combined with automated remediation actions considerably reduces the human effort. We look at monitoring data in a big provider's domain and abstract out the repeatable tasks from the noisy and unstructured human-readable text in tickets. We present a novel approach for automatic problem determination from this noisy and unstructured text. The approach uses two distinct levels of analysis, (a) correlating different data sources to obtain a richer text followed by (b) context based classification of the correlated data. We report on accuracy and efficiency of our approach using real customer data.

This paper presents a novel approach to categorize, model and identify contextual information in natural language interface to database (NLIDB) systems. The interactions between user and system are categorized and modeled based on the way in which the contextual information is utilized in the interactions. A relationship schema among the responses (user and system responses) is proposed. We present a novel method to identify contextual information in one specific type of usersystem interaction. We report on results of experiments with the university related queries.

This paper presents a novel approach to building natural language interface to databases (NLIDB) based on Computational Paninian Grammar (CPG). It uses two distinct stages of processing, namely, syntactic processing followed by semantic processing. Syntactic processing makes the processing more general and robust. CPG is a dependency framework in which the analysis is in terms of syntacticosemantic relations. The closeness of these relations makes semantic processing easier and more accurate. It also makes the systems more portable.