Search for: All records

The events in a narrative are understood as a coherent whole via the underlying states of their participants. Often, these participant states are not explicitly mentioned, instead left to be inferred by the reader. A model that understands narratives should likewise infer these implicit states, and even reason about the impact of changes to these states on the narrative. To facilitate this goal, we introduce a new crowdsourced English-language, Participant States dataset, PASTA. This dataset contains inferable participant states; a counterfactual perturbation to each state; and the changes to the story that would be necessary if the counterfactual were true. We introduce three state-based reasoning tasks that test for the ability to infer when a state is entailed by a story, to revise a story conditioned on a counterfactual state, and to explain the most likely state change given a revised story. Experiments show that today’s LLMs can reason about states to some degree, but there is large room for improvement, especially in problems requiring access and ability to reason with diverse types of knowledge (e.g. physical, numerical, factual). 

The proliferation of Large Language Models (LLMs) presents both a critical design challenge and a remarkable opportunity for the field of Human-Robot Interaction (HRI). While the direct deployment of LLMs on interactive robots may be unsuitable for reasons of ethics, safety, and control, LLMs might nevertheless provide a promising baseline technique for many elements of HRI. Specifically, in this position paper, we argue for the use of LLMs as Scarecrows: ‘brainless,’ straw-man black-box modules integrated into robot architectures for the purpose of quickly enabling full-pipeline solutions, much like the use of “Wizard of Oz” (WoZ) and other human-in-the-loop approaches. We explicitly acknowledge that these Scarecrows, rather than providing a satisfying or scientifically complete solution, incorporate a form of the wisdom of the crowd, and, in at least some cases, will ultimately need to be replaced or supplemented by a robust and theoretically motivated solution. We provide examples of how Scarecrows could be used in language-capable robot architectures as useful placeholders, and suggest initial reporting guidelines for authors, mirroring existing guidelines for the use and reporting of WoZ techniques. 

Over the past decade, the machine learning security community has developed a myriad of defenses for evasion attacks. An under- studied question in that community is: for whom do these defenses defend? This work considers common approaches to defending learned systems and how security defenses result in performance inequities across different sub-populations. We outline appropriate parity metrics for analysis and begin to answer this question through empirical results of the fairness implications of machine learning security methods. We find that many methods that have been proposed can cause direct harm, like false rejection and unequal benefits from robustness training. The framework we propose for measuring defense equality can be applied to robustly trained models, preprocessing-based defenses, and rejection methods. We identify a set of datasets with a user-centered application and a reasonable computational cost suitable for case studies in measuring the equality of defenses. In our case study of speech command recognition, we show how such adversarial training and augmentation have non-equal but complex protections for social subgroups across gender, accent, and age in relation to user coverage. We present a comparison of equality between two rejection-based de- fenses: randomized smoothing and neural rejection, finding randomized smoothing more equitable due to the sampling mechanism for minority groups. This represents the first work examining the disparity in the adversarial robustness in the speech domain and the fairness evaluation of rejection-based defenses. 

Prior work has shown that coupling sequential latent variable models with semantic ontological knowledge can improve the representational capabilities of event modeling ap- proaches. In this work, we present a novel, doubly hierarchical, semi-supervised event modeling framework that provides structural hierarchy while also accounting for ontological hierarchy. Our approach consists of multiple layers of structured latent variables, where each successive layer compresses and abstracts the previous layers. We guide this compression through the injection of structured ontological knowledge that is defined at the type level of events: importantly, our model allows for partial injection of semantic knowledge and it does not depend on observing instances at any particular level of the semantic ontology. Across two different datasets and four different evaluation metrics, we demonstrate that our approach is able to out-perform the previous state-of-the-art approaches by up to 8.5%, demonstrating the benefits of structured and semantic hierarchical knowledge for event modeling. 

Language models have shown great promise in common-sense related tasks. However, it remains unseen how they would perform in the context of physically situated human-robot interactions, particularly in disaster-relief scenarios. In this paper, we develop a language model evaluation dataset with more than 800 cloze sentences, written to probe for the function of over 200 objects. The sentences are divided into two tasks: an “easy” task where the language model has to choose between vocabulary with different functions (Task 1), and a “challenge” where it has to choose between vocabulary with the same function, yet only one vocabulary item is appropriate given real world constraints on functionality (Task 2). DistilBERT performs with about 80% accuracy for both tasks. To investigate how annotator variability affected those results, we developed a follow-on experiment where we compared our original results with wrong answers chosen based on embedding vector distances. Those results showed increased precision across documents but a 15% decrease in accuracy. We conclude that language models do have a strong knowledge basis for object reasoning, but will require creative fine-tuning strategies in order to be successfully deployed. 

The existence of external (“side”) semantic knowledge has been shown to result in more expressive computational event models. To enable the use of side information that may be noisy or missing, we propose a semi-supervised information bottleneck-based discrete latent variable model. We reparameterize the model’s discrete variables with auxiliary continuous latent variables and a light-weight hierarchical structure. Our model is learned to minimize the mutual information between the observed data and optional side knowledge that is not already captured by the new, auxiliary variables. We theoretically show that our approach generalizes past approaches, and perform an empirical case study of our approach on event modeling. We corroborate our theoretical results with strong empirical experiments, showing that the proposed method outperforms previous proposed approaches on multiple datasets. 

Many metric learning tasks, such as triplet learning, nearest neighbor retrieval, and visualization, are treated primarily as embedding tasks where the ultimate metric is some variant of the Euclidean distance (e.g., cosine or Mahalanobis), and the algorithm must learn to embed points into the pre-chosen space. The study of non-Euclidean geometries is often not explored, which we believe is due to a lack of tools for learning non-Euclidean measures of distance. Recent work has shown that Bregman divergences can be learned from data, opening a promising approach to learning asymmetric distances. We propose a new approach to learning arbitrary Bergman divergences in a differentiable manner via input convex neural networks and show that it overcomes significant limitations of previous works. We also demonstrate that our method more faithfully learns divergences over a set of both new and previously studied tasks, including asymmetric regression, ranking, and clustering. Our tests further extend to known asymmetric, but non-Bregman tasks, where our method still performs competitively despite misspecification, showing the general utility of our approach for asymmetric learning. 

In this paper, we present a shared manipulation task performed both in virtual reality with a simulated robot and in the real world with a physical robot. A collaborative assembly task where the human and robot work together to construct as simple electrical circuit was chosen. While there are platforms available for conducting human robot interactions using virtual reality, there has not been significant work investigating how it can influence human perception of tasks that are typically done in person. We present an overview of the simulation environment used, describe the paired experiment being performed, and finally enumerate a set of design desiderata to be considered when conducting sim2real experiment involving humans in a virtual setting. 

Machine learning models that sense human speech, body placement, and other key features are commonplace in human-robot interaction. However, the deployment of such models in themselves is not without risk. Research in the security of machine learning examines how such models can be exploited and the risks associated with these exploits. Unfortunately, the threat models of risks produced by machine learning security do not incorporate the rich sociotechnical underpinnings of the defenses they propose; as a result, efforts to improve the security of machine learning models may actually increase the difference in performance across different demographic groups, yielding systems that have risk mitigation that work better for one group than another. In this work, we outline why current approaches to machine learning security present DEI concerns for the human-robot interaction community and where there are open areas for collaboration.