Multiple Distribution Shift -- Aerial (MDS-A): A Dataset for Test-Time Error Detection and Model Adaptation

Machine learning models assume that training and test samples are drawn from the same distribution. As such, significant differences between training and test distributions often lead to degradations in performance. We introduce Multiple Distribution Shift -- Aerial (MDS-A) -- a collection of inter-related datasets of the same aerial domain that are perturbed in different ways to better characterize the effects of out-of-distribution performance. Specifically, MDS-A is a set of simulated aerial datasets collected under different weather conditions. We include six datasets under different simulated weather conditions along with six baseline object-detection models, as well as several test datasets that are a mix of weather conditions that we show have significant differences from the training data. In this paper, we present characterizations of MDS-A, provide performance results for the baseline machine learning models (on both their specific training datasets and the test data), as well as results of the baselines after employing recent knowledge-engineering error-detection techniques (EDR) thought to improve out-of-distribution performance. The dataset is available at https://lab-v2.github.io/mdsa-dataset-website.

MoSH: Modeling Multi-Objective Tradeoffs with Soft and Hard Bounds

Countless science and engineering applications in multi-objective optimization (MOO) necessitate that decision-makers (DMs) select a Pareto-optimal solution which aligns with their preferences. Evaluating individual solutions is often expensive, necessitating cost-sensitive optimization techniques. Due to competing objectives, the space of trade-offs is also expansive -- thus, examining the full Pareto frontier may prove overwhelming to a DM. Such real-world settings generally have loosely-defined and context-specific desirable regions for each objective function that can aid in constraining the search over the Pareto frontier. We introduce a novel conceptual framework that operationalizes these priors using soft-hard functions, SHFs, which allow for the DM to intuitively impose soft and hard bounds on each objective -- which has been lacking in previous MOO frameworks. Leveraging a novel minimax formulation for Pareto frontier sampling, we propose a two-step process for obtaining a compact set of Pareto-optimal points which respect the user-defined soft and hard bounds: (1) densely sample the Pareto frontier using Bayesian optimization, and (2) sparsify the selected set to surface to the user, using robust submodular function optimization. We prove that (2) obtains the optimal compact Pareto-optimal set of points from (1). We further show that many practical problems fit within the SHF framework and provide extensive empirical validation on diverse domains, including brachytherapy, engineering design, and large language model personalization. Specifically, for brachytherapy, our approach returns a compact set of points with over 3% greater SHF-defined utility than the next best approach. Among the other diverse experiments, our approach consistently leads in utility, allowing the DM to reach >99% of their maximum possible desired utility within validation of 5 points.

RoboFail: Analyzing Failures in Robot Learning Policies

Despite being trained on increasingly large datasets, robot models often overfit to specific environments or datasets. Consequently, they excel within their training distribution but face challenges in generalizing to novel or unforeseen scenarios. This paper presents a method to proactively identify failure mode probabilities in robot manipulation policies, providing insights into where these models are likely to falter. To this end, since exhaustively searching over a large space of failures is infeasible, we propose a deep reinforcement learning-based framework, RoboFail. It is designed to detect scenarios prone to failure and quantify their likelihood, thus offering a structured approach to anticipate failures. By identifying these high-risk states in advance, RoboFail enables researchers and engineers to better understand the robustness limits of robot policies, contributing to the development of safer and more adaptable robotic systems.

ExpressivityArena: Can LLMs Express Information Implicitly?

While Large Language Models (LLMs) have demonstrated remarkable performance in certain dimensions, their ability to express implicit language cues that human use for effective communication remains unclear. This paper presents ExpressivityArena, a Python library for measuring the implicit communication abilities of LLMs. We provide a comprehensive framework to evaluate expressivity of arbitrary LLMs and explore its practical implications. To this end, we refine the definition and measurements of ``expressivity,'' and use our framework in a set of small experiments. These experiments test LLMs in creative and logical tasks such as poetry, coding, and emotion-based responses. They are then evaluated by an automated grader, through ExpressivityArena, which we verify to be the most pragmatic for testing expressivity. Building on these experiments, we deepen our understanding of the expressivity of LLMs by assessing their ability to remain expressive in conversations. Our findings indicate that LLMs are capable of generating and understanding expressive content, however, with some limitations. These insights will inform the future development and deployment of expressive LLMs. We provide the code for ExpressivityArena alongside our paper.

LLM-Assisted Red Teaming of Diffusion Models through "Failures Are Fated, But Can Be Faded"

In large deep neural networks that seem to perform surprisingly well on many tasks, we also observe a few failures related to accuracy, social biases, and alignment with human values, among others. Therefore, before deploying these models, it is crucial to characterize this failure landscape for engineers to debug or audit models. Nevertheless, it is infeasible to exhaustively test for all possible combinations of factors that could lead to a model's failure. In this paper, we improve the "Failures are fated, but can be faded" framework (arXiv:2406.07145)--a post-hoc method to explore and construct the failure landscape in pre-trained generative models--with a variety of deep reinforcement learning algorithms, screening tests, and LLM-based rewards and state generation. With the aid of limited human feedback, we then demonstrate how to restructure the failure landscape to be more desirable by moving away from the discovered failure modes. We empirically demonstrate the effectiveness of the proposed method on diffusion models. We also highlight the strengths and weaknesses of each algorithm in identifying failure modes.

Trustworthy Conceptual Explanations for Neural Networks in Robot Decision-Making

Black box neural networks are an indispensable part of modern robots. Nevertheless, deploying such high-stakes systems in real-world scenarios poses significant challenges when the stakeholders, such as engineers and legislative bodies, lack insights into the neural networks' decision-making process. Presently, explainable AI is primarily tailored to natural language processing and computer vision, falling short in two critical aspects when applied in robots: grounding in decision-making tasks and the ability to assess trustworthiness of their explanations. In this paper, we introduce a trustworthy explainable robotics technique based on human-interpretable, high-level concepts that attribute to the decisions made by the neural network. Our proposed technique provides explanations with associated uncertainty scores by matching neural network's activations with human-interpretable visualizations. To validate our approach, we conducted a series of experiments with various simulated and real-world robot decision-making models, demonstrating the effectiveness of the proposed approach as a post-hoc, human-friendly robot learning diagnostic tool.

Explainable Concept Generation through Vision-Language Preference Learning

Concept-based explanations have become a popular choice for explaining deep neural networks post-hoc because, unlike most other explainable AI techniques, they can be used to test high-level visual "concepts" that are not directly related to feature attributes. For instance, the concept of "stripes" is important to classify an image as a zebra. Concept-based explanation methods, however, require practitioners to guess and collect multiple candidate concept image sets, which can often be imprecise and labor-intensive. Addressing this limitation, in this paper, we frame concept image set creation as an image generation problem. However, since naively using a generative model does not result in meaningful concepts, we devise a reinforcement learning-based preference optimization algorithm that fine-tunes the vision-language generative model from approximate textual descriptions of concepts. Through a series of experiments, we demonstrate the capability of our method to articulate complex, abstract concepts that are otherwise challenging to craft manually. In addition to showing the efficacy and reliability of our method, we show how our method can be used as a diagnostic tool for analyzing neural networks.

Towards Adapting Reinforcement Learning Agents to New Tasks: Insights from Q-Values

While contemporary reinforcement learning research and applications have embraced policy gradient methods as the panacea of solving learning problems, value-based methods can still be useful in many domains as long as we can wrangle with how to exploit them in a sample efficient way. In this paper, we explore the chaotic nature of DQNs in reinforcement learning, while understanding how the information that they retain when trained can be repurposed for adapting a model to different tasks. We start by designing a simple experiment in which we are able to observe the Q-values for each state and action in an environment. Then we train in eight different ways to explore how these training algorithms affect the way that accurate Q-values are learned (or not learned). We tested the adaptability of each trained model when retrained to accomplish a slightly modified task. We then scaled our setup to test the larger problem of an autonomous vehicle at an unprotected intersection. We observed that the model is able to adapt to new tasks quicker when the base model's Q-value estimates are closer to the true Q-values. The results provide some insights and guidelines into what algorithms are useful for sample efficient task adaptation.

Failures Are Fated, But Can Be Faded: Characterizing and Mitigating Unwanted Behaviors in Large-Scale Vision and Language Models

In large deep neural networks that seem to perform surprisingly well on many tasks, we also observe a few failures related to accuracy, social biases, and alignment with human values, among others. Therefore, before deploying these models, it is crucial to characterize this failure landscape for engineers to debug and legislative bodies to audit models. Nevertheless, it is infeasible to exhaustively test for all possible combinations of factors that could lead to a model's failure. In this paper, we introduce a post-hoc method that utilizes \emph{deep reinforcement learning} to explore and construct the landscape of failure modes in pre-trained discriminative and generative models. With the aid of limited human feedback, we then demonstrate how to restructure the failure landscape to be more desirable by moving away from the discovered failure modes. We empirically show the effectiveness of the proposed method across common Computer Vision, Natural Language Processing, and Vision-Language tasks.

Fairness in Autonomous Driving: Towards Understanding Confounding Factors in Object Detection under Challenging Weather

The deployment of autonomous vehicles (AVs) is rapidly expanding to numerous cities. At the heart of AVs, the object detection module assumes a paramount role, directly influencing all downstream decision-making tasks by considering the presence of nearby pedestrians, vehicles, and more. Despite high accuracy of pedestrians detected on held-out datasets, the potential presence of algorithmic bias in such object detectors, particularly in challenging weather conditions, remains unclear. This study provides a comprehensive empirical analysis of fairness in detecting pedestrians in a state-of-the-art transformer-based object detector. In addition to classical metrics, we introduce novel probability-based metrics to measure various intricate properties of object detection. Leveraging the state-of-the-art FACET dataset and the Carla high-fidelity vehicle simulator, our analysis explores the effect of protected attributes such as gender, skin tone, and body size on object detection performance in varying environmental conditions such as ambient darkness and fog. Our quantitative analysis reveals how the previously overlooked yet intuitive factors, such as the distribution of demographic groups in the scene, the severity of weather, the pedestrians' proximity to the AV, among others, affect object detection performance. Our code is available at https://github.com/bimsarapathiraja/fair-AV.