Research Papers

Responsible AI has risen to the forefront of the AI research community. As neural network-based learning algorithms continue to permeate real-world applications, the field of Responsible AI has played a large role in ensuring that such systems maintain a high-level of human-compatibility. Despite this progress, the state of the art in Responsible AI has ignored one crucial point: human problems are multi-agent problems. Predominant approaches largely consider the performance of a single AI system in isolation, but human problems are, by their very nature, multi-agent. From driving in traffic to negotiating economic policy, human problem-solving involves interaction and the interplay of the actions and motives of multiple individuals. This dissertation develops the study of responsible emergent multi-agent behavior, illustrating how researchers and practitioners can better understand and shape multi-agent learning with respect to three pillars of Responsible AI: interpretability, fairness, and robustness. First, I investigate multi-agent interpretability, presenting novel techniques for understanding emergent multi-agent behavior at multiple levels of granularity. With respect to low-level interpretability, I examine the extent to which implicit communication emerges as an aid to coordination in multi-agent populations. I introduce a novel curriculum-driven method for learning high-performing policies in difficult, sparse reward environments and show through a measure of position-based social influence that multi-agent teams that learn sophisticated coordination strategies exchange significantly more information through implicit signals than lesser-coordinated agents. Then, at a high-level, I study concept-based interpretability in the context of multi-agent learning. I propose a novel method for learning intrinsically interpretable, concept-based policies and show that it enables...

Exploring Practitioner Perspectives On Training Data Attribution Explanations

Elisa Nguyen, Evgenii Kortukov, Jean Y. Song +1

Explainable AI (XAI) aims to provide insight into opaque model reasoning to humans and as such is an interdisciplinary field by nature. In this paper, we interviewed 10 practitioners to understand the possible usability of training data attribution (TDA) explanations and to explore the design space of such an approach. We confirmed that training data quality is often the most important factor for high model performance in practice and model developers mainly rely on their own experience to curate data. End-users expect explanations to enhance their interaction with the model and do not necessarily prioritise but are open to training data as a means of explanation. Within our participants, we found that TDA explanations are not well-known and therefore not used. We urge the community to focus on the utility of TDA techniques from the human-machine collaboration perspective and broaden the TDA evaluation to reflect common use cases in practice.

Educational disparities within the Dominican Republic (DR) have long-standing origins rooted in economic, political, and social inequity. Addressing these challenges has necessarily called for capacity building with respect to educational materials, high-quality instruction, and structural resourcing. Generative AI tools like ChatGPT have begun to pique the interest of Dominican educators due to their perceived potential to bridge these educational gaps. However, a substantial body of AI fairness literature has documented ways AI disproportionately reinforces power dynamics reflective of jurisdictions driving AI development and deployment policies, collectively termed the AI Global North. As such, indiscriminate adoption of this technology for DR education, even in part, risks perpetuating forms of digital coloniality. Therefore, this paper centers embracing AI-facilitated educational reform by critically examining how AI-driven tools like ChatGPT in DR education may replicate facets of digital colonialism. We provide a concise overview of 20th-century Dominican education reforms following the 1916 US occupation. Then, we employ identified neocolonial aspects historically shaping Dominican education to interrogate the perceived advantages of ChatGPT for contemporary Dominican education, as outlined by a Dominican scholar. This work invites AI Global North & South developers, stakeholders, and Dominican leaders alike to exercise a relational contextualization of data-centric epistemologies like ChatGPT to reap its transformative benefits while remaining vigilant of safeguarding Dominican digital sovereignty.

Explainable artificial intelligence (XAI) is one of the most intensively developed area of AI in recent years. It is also one of the most fragmented with multiple methods that focus on different aspects of explanations. This makes difficult to obtain the full spectrum of explanation at once in a compact and consistent way. To address this issue, we present Local Universal Explainer (LUX), which is a rule-based explainer that can generate factual, counterfactual and visual explanations. It is based on a modified version of decision tree algorithms that allows for oblique splits and integration with feature importance XAI methods such as SHAP. It limits the use data generation in opposite to other algorithms, but is focused on selecting local concepts in a form of high-density clusters of real data that have the highest impact on forming the decision boundary of the explained model and generating artificial samples with novel SHAP-guided sampling algorithm. We tested our method on real and synthetic datasets and compared it with state-of-the-art rule-based explainers such as LORE, EXPLAN and Anchor. Our method outperforms the existing approaches in terms of simplicity, fidelity, representativeness, and consistency.

Explainable AI (XAI) is the study on how humans can be able to understand the cause of a model's prediction. In this work, the problem of interest is Scene Text Recognition (STR) Explainability, using XAI to understand the cause of an STR model's prediction. Recent XAI literatures on STR only provide a simple analysis and do not fully explore other XAI methods. In this study, we specifically work on data explainability frameworks, called attribution-based methods, that explain the important parts of an input data in deep learning models. However, integrating them into STR produces inconsistent and ineffective explanations, because they only explain the model in the global context. To solve this problem, we propose a new method, STRExp, to take into consideration the local explanations, i.e. the individual character prediction explanations. This is then benchmarked across different attribution-based methods on different STR datasets and evaluated across different STR models.

SurroCBM: Concept Bottleneck Surrogate Models for Generative Post-hoc Explanation

Bo Pan, Zhenke Liu, Yifei Zhang +1

Explainable AI seeks to bring light to the decision-making processes of black-box models. Traditional saliency-based methods, while highlighting influential data segments, often lack semantic understanding. Recent advancements, such as Concept Activation Vectors (CAVs) and Concept Bottleneck Models (CBMs), offer concept-based explanations but necessitate human-defined concepts. However, human-annotated concepts are expensive to attain. This paper introduces the Concept Bottleneck Surrogate Models (SurroCBM), a novel framework that aims to explain the black-box models with automatically discovered concepts. SurroCBM identifies shared and unique concepts across various black-box models and employs an explainable surrogate model for post-hoc explanations. An effective training strategy using self-generated data is proposed to enhance explanation quality continuously. Through extensive experiments, we demonstrate the efficacy of SurroCBM in concept discovery and explanation, underscoring its potential in advancing the field of explainable AI.

We propose an approach for early crop classification through identifying important timesteps with eXplainable AI (XAI) methods. Our approach consists of training a baseline crop classification model to carry out layer-wise relevance propagation (LRP) so that the salient time step can be identified. We chose a selected number of such important time indices to create the bounding region of the shortest possible classification timeframe. We identified the period 21st April 2019 to 9th August 2019 as having the best trade-off in terms of accuracy and earliness. This timeframe only suffers a 0.75% loss in accuracy as compared to using the full timeseries. We observed that the LRP-derived important timesteps also highlight small details in input values that differentiates between different classes and

The ability to navigate robots with natural language instructions in an unknown environment is a crucial step for achieving embodied artificial intelligence (AI). With the improving performance of deep neural models proposed in the field of vision-and-language navigation (VLN), it is equally interesting to know what information the models utilize for their decision-making in the navigation tasks. To understand the inner workings of deep neural models, various explanation methods have been developed for promoting explainable AI (XAI). But they are mostly applied to deep neural models for image or text classification tasks and little work has been done in explaining deep neural models for VLN tasks. In this paper, we address these problems by building quantitative benchmarks to evaluate explanation methods for VLN models in terms of faithfulness. We propose a new erasure-based evaluation pipeline to measure the step-wise textual explanation in the sequential decision-making setting. We evaluate several explanation methods for two representative VLN models on two popular VLN datasets and reveal valuable findings through our experiments.

Black-box Attacks on Image Activity Prediction and its Natural Language Explanations

Alina Elena Baia, Valentina Poggioni, Andrea Cavallaro

Explainable AI (XAI) methods aim to describe the decision process of deep neural networks. Early XAI methods produced visual explanations, whereas more recent techniques generate multimodal explanations that include textual information and visual representations. Visual XAI methods have been shown to be vulnerable to white-box and gray-box adversarial attacks, with an attacker having full or partial knowledge of and access to the target system. As the vulnerabilities of multimodal XAI models have not been examined, in this paper we assess for the first time the robustness to black-box attacks of the natural language explanations generated by a self-rationalizing image-based activity recognition model. We generate unrestricted, spatially variant perturbations that disrupt the association between the predictions and the corresponding explanations to mislead the model into generating unfaithful explanations. We show that we can create adversarial images that manipulate the explanations of an activity recognition model by having access only to its final output.

On the Trade-offs between Adversarial Robustness and Actionable Explanations

Satyapriya Krishna, Chirag Agarwal, Himabindu Lakkaraju

As machine learning models are increasingly being employed in various high-stakes settings, it becomes important to ensure that predictions of these models are not only adversarially robust, but also readily explainable to relevant stakeholders. However, it is unclear if these two notions can be simultaneously achieved or if there exist trade-offs between them. In this work, we make one of the first attempts at studying the impact of adversarially robust models on actionable explanations which provide end users with a means for recourse. We theoretically and empirically analyze the cost (ease of implementation) and validity (probability of obtaining a positive model prediction) of recourses output by state-of-the-art algorithms when the underlying models are adversarially robust vs. non-robust. More specifically, we derive theoretical bounds on the differences between the cost and the validity of the recourses generated by state-of-the-art algorithms for adversarially robust vs. non-robust linear and non-linear models. Our empirical results with multiple real-world datasets validate our theoretical results and show the impact of varying degrees of model robustness on the cost and validity of the resulting recourses. Our analyses demonstrate that adversarially robust models significantly increase the cost and reduce the validity of the resulting recourses, thus shedding light on the inherent trade-offs between adversarial robustness and actionable explanations.

Explaining Deep Face Algorithms through Visualization: A Survey

Thrupthi Ann John, Vineeth N Balasubramanian, C. V. Jawahar

Although current deep models for face tasks surpass human performance on some benchmarks, we do not understand how they work. Thus, we cannot predict how it will react to novel inputs, resulting in catastrophic failures and unwanted biases in the algorithms. Explainable AI helps bridge the gap, but currently, there are very few visualization algorithms designed for faces. This work undertakes a first-of-its-kind meta-analysis of explainability algorithms in the face domain. We explore the nuances and caveats of adapting general-purpose visualization algorithms to the face domain, illustrated by computing visualizations on popular face models. We review existing face explainability works and reveal valuable insights into the structure and hierarchy of face networks. We also determine the design considerations for practical face visualizations accessible to AI practitioners by conducting a user study on the utility of various explainability algorithms.

Explainable Artificial Intelligence (XAI) models have recently attracted a great deal of interest from a variety of application sectors. Despite significant developments in this area, there are still no standardized methods or approaches for understanding AI model outputs. A systematic and cohesive framework is also increasingly necessary to incorporate new techniques like discriminative and generative models to close the gap. This paper contributes to the discourse on XAI by presenting an empirical evaluation based on a novel framework: Sampling - Variational Auto Encoder (VAE) - Ensemble Anomaly Detection (SVEAD). It is a hybrid architecture where VAE combined with ensemble stacking and SHapley Additive exPlanations are used for imbalanced classification. The finding reveals that combining ensemble stacking, VAE, and SHAP can. not only lead to better model performance but also provide an easily explainable framework. This work has used SHAP combined with Permutation Importance and Individual Conditional Expectations to create a powerful interpretability of the model. The finding has an important implication in the real world, where the need for XAI is paramount to boost confidence in AI applications.