Total: 12
With Artificial Intelligence on the rise, human interaction with autonomous agents becomes more frequent. Effective human-agent collaboration requires users to understand the agent's behavior, as failing to do so may cause reduced productivity, misuse or frustration. Agent strategy summarization methods are used to describe the strategy of an agent to users through demonstrations. A summary's objective is to maximize the user's understanding of the agent's aptitude by showcasing its behaviour in a selected set of world states. While shown to be useful, we show that current methods are limited when tasked with comparing between agents, as each summary is independently generated for a specific agent. In this paper, we propose a novel method for generating dependent and contrastive summaries that emphasize the differences between agent policies by identifying states in which the agents disagree on the best course of action. We conducted user studies to assess the usefulness of disagreement-based summaries for identifying superior agents and conveying agent differences. Results show disagreement-based summaries lead to improved user performance compared to summaries generated using HIGHLIGHTS, a strategy summarization algorithm which generates summaries for each agent independently.
In attempts to "explain" predictions of machine learning models, researchers have proposed hundreds of techniques for attributing predictions to features that are deemed important. While these attributions are often claimed to hold the potential to improve human "understanding" of the models, surprisingly little work explicitly evaluates progress towards this aspiration. In this paper, we conduct a crowdsourcing study, where participants interact with deception detection models that have been trained to distinguish between genuine and fake hotel reviews. They are challenged both to simulate the model on fresh reviews, and to edit reviews with the goal of lowering the probability of the originally predicted class. Successful manipulations would lead to an adversarial example. During the training (but not the test) phase, input spans are highlighted to communicate salience. Through our evaluation, we observe that for a linear bag-of-words model, participants with access to the feature coefficients during training are able to cause a larger reduction in model confidence in the testing phase when compared to the no-explanation control. For the BERT-based classifier, popular local explanations do not improve their ability to reduce the model confidence over the no-explanation case. Remarkably, when the explanation for the BERT model is given by the (global) attributions of a linear model trained to imitate the BERT model, people can effectively manipulate the model.
Recent work has shown the potential benefit of selective prediction systems that can learn to defer to a human when the predictions of the AI are unreliable, particularly to improve the reliability of AI systems in high-stakes applications like healthcare or conservation. However, most prior work assumes that human behavior remains unchanged when they solve a prediction task as part of a human-AI team as opposed to by themselves. We show that this is not the case by performing experiments to quantify human-AI interaction in the context of selective prediction. In particular, we study the impact of communicating different types of information to humans about the AI system's decision to defer. Using real-world conservation data and a selective prediction system that improves expected accuracy over that of the human or AI system working individually, we show that this messaging has a significant impact on the accuracy of human judgements. Our results study two components of the messaging strategy: 1) Whether humans are informed about the prediction of the AI system and 2) Whether they are informed about the decision of the selective prediction system to defer. By manipulating these messaging components, we show that it is possible to significantly boost human performance by informing the human of the decision to defer, but not revealing the prediction of the AI. We therefore show that it is vital to consider how the decision to defer is communicated to a human when designing selective prediction systems, and that the composite accuracy of a human-AI team must be carefully evaluated using a human-in-the-loop framework.
The new generation of language models is reported to solve some extraordinary tasks the models were never trained for specifically, in few-shot or zero-shot settings. However, these reports usually cherry-pick the tasks, use the best prompts, and unwrap or extract the solutions leniently even if they are followed by nonsensical text. In sum, they are specialised results for one domain, a particular way of using the models and interpreting the results. In this paper, we present a novel theoretical evaluation framework and a distinctive experimental study assessing language models as general-purpose systems when used directly by human prompters --- in the wild. For a useful and safe interaction in these increasingly more common conditions, we need to understand when the model fails because of a lack of capability or a misunderstanding of the user's intents. Our results indicate that language models such as GPT-3 have limited understanding of the human command; far from becoming general-purpose systems in the wild.
Learning from Crowds (LFC) seeks to induce a high-quality classifier from training instances, which are linked to a range of possible noisy annotations from crowdsourcing workers under their various levels of skills and their own preconditions. Recent studies on LFC focus on designing new methods to improve the performance of the classifier trained from crowdsourced labeled data. To this day, however, there remain under-explored security aspects of LFC systems. In this work, we seek to bridge this gap. We first show that LFC models are vulnerable to adversarial examples---small changes to input data can cause classifiers to make prediction mistakes. Second, we propose an approach, A-LFC for training a robust classifier from crowdsourced labeled data. Our empirical results on three real-world datasets show that the proposed approach can substantially improve the performance of the trained classifier even with the existence of adversarial examples. On average, A-LFC has 10.05% and 11.34% higher test robustness than the state-of-the-art in the white-box and black-box attack settings, respectively.
Model interpretability has become an important problem in machine learning (ML) due to the increased effect algorithmic decisions have on humans. Counterfactual explanations can help users understand not only why ML models make certain decisions, but also how these decisions can be changed. We frame the problem of finding counterfactual explanations as an optimization task and extend previous work that could only be applied to differentiable models. In order to accommodate non-differentiable models such as tree ensembles, we use probabilistic model approximations in the optimization framework. We introduce an approximation technique that is effective for finding counterfactual explanations for predictions of the original model and show that our counterfactual examples are significantly closer to the original instances than those produced by other methods specifically designed for tree ensembles.
Expert decision makers are starting to rely on data-driven automated agents to assist them with various tasks. For this collaboration to perform properly, the human decision maker must have a mental model of when and when not to rely on the agent. In this work, we aim to ensure that human decision makers learn a valid mental model of the agent's strengths and weaknesses. To accomplish this goal, we propose an exemplar-based teaching strategy where humans solve a set of selected examples and with our help generalize from them to the domain. We present a novel parameterization of the human's mental model of the AI that applies a nearest neighbor rule in local regions surrounding the teaching examples. Using this model, we derive a near-optimal strategy for selecting a representative teaching set. We validate the benefits of our teaching strategy on a multi-hop question answering task with an interpretable AI model using crowd workers. We find that when workers draw the right lessons from the teaching stage, their task performance improves. We furthermore validate our method on a set of synthetic experiments.
We study the design of autonomous agents that are capable of deceiving outside observers about their intentions while carrying out tasks in stochastic, complex environments. By modeling the agent's behavior as a Markov decision process, we consider a setting where the agent aims to reach one of multiple potential goals while deceiving outside observers about its true goal. We propose a novel approach to model observer predictions based on the principle of maximum entropy and to efficiently generate deceptive strategies via linear programming. The proposed approach enables the agent to exhibit a variety of tunable deceptive behaviors while ensuring the satisfaction of probabilistic constraints on the behavior. We evaluate the performance of the proposed approach via comparative user studies and present a case study on the streets of Manhattan, New York, using real travel time distributions.
Shared autonomy refers to approaches for enabling an autonomous agent to collaborate with a human with the aim of improving human performance. However, besides improving performance, it may often also be beneficial that the agent concurrently accounts for preserving the user’s experience or satisfaction of collaboration. In order to address this additional goal, we examine approaches for improving the user experience by constraining the number of interventions by the autonomous agent. We propose two model-free reinforcement learning methods that can account for both hard and soft constraints on the number of interventions. We show that not only does our method outperform the existing baseline, but also eliminates the need to manually tune a black-box hyperparameter for controlling the level of assistance. We also provide an in-depth analysis of intervention scenarios in order to further illuminate system understanding.
State-of-the-art brain-to-text systems have achieved great success in decoding language directly from brain signals using neural networks. However, current approaches are limited to small closed vocabularies which are far from enough for natural communication. In addition, most of the high-performing approaches require data from invasive devices (e.g., ECoG). In this paper, we extend the problem to open vocabulary Electroencephalography(EEG)-To-Text Sequence-To-Sequence decoding and zero-shot sentence sentiment classification on natural reading tasks. We hypothesis that the human brain functions as a special text encoder and propose a novel framework leveraging pre-trained language models (e.g., BART). Our model achieves a 40.1% BLEU-1 score on EEG-To-Text decoding and a 55.6% F1 score on zero-shot EEG-based ternary sentiment classification, which significantly outperforms supervised baselines. Furthermore, we show that our proposed model can handle data from various subjects and sources, showing great potential for a high-performance open vocabulary brain-to-text system once sufficient data is available. The code is made publicly available for research purpose at https://github.com/MikeWangWZHL/EEG-To-Text.
Understanding how the predictions of deep learning models are formed during the training process is crucial to improve model performance and fix model defects, especially when we need to investigate nontrivial training strategies such as active learning, and track the root cause of unexpected training results such as performance degeneration. In this work, we propose a time-travelling visual solution DeepVisualInsight (DVI), aiming to manifest the spatio-temporal causality while training a deep learning image classifier. The spatio-temporal causality demonstrates how the gradient-descent algorithm and various training data sampling techniques can influence and reshape the layout of learnt input representation and the classification boundaries in consecutive epochs. Such causality allows us to observe and analyze the whole learning process in the visible low dimensional space. Technically, we propose four spatial and temporal properties and design our visualization solution to satisfy them. These properties preserve the most important information when projecting and inverse-projecting input samples between the visible low-dimensional and the invisible high-dimensional space, for causal analyses. Our extensive experiments show that, comparing to baseline approaches, we achieve the best visualization performance regarding the spatial/temporal properties and visualization efficiency. Moreover, our case study shows that our visual solution can well reflect the characteristics of various training scenarios, showing good potential of DVI as a debugging tool for analyzing deep learning training processes.
Human emotions involve basic and compound facial expressions. However, current research on facial expression recognition (FER) mainly focuses on basic expressions, and thus fails to address the diversity of human emotions in practical scenarios. Meanwhile, existing work on compound FER relies heavily on abundant labeled compound expression training data, which are often laboriously collected under the professional instruction of psychology. In this paper, we study compound FER in the cross-domain few-shot learning setting, where only a few images of novel classes from the target domain are required as a reference. In particular, we aim to identify unseen compound expressions with the model trained on easily accessible basic expression datasets. To alleviate the problem of limited base classes in our FER task, we propose a novel Emotion Guided Similarity Network (EGS-Net), consisting of an emotion branch and a similarity branch, based on a two-stage learning framework. Specifically, in the first stage, the similarity branch is jointly trained with the emotion branch in a multi-task fashion. With the regularization of the emotion branch, we prevent the similarity branch from overfitting to sampled base classes that are highly overlapped across different episodes. In the second stage, the emotion branch and the similarity branch play a “two-student game” to alternately learn from each other, thereby further improving the inference ability of the similarity branch on unseen compound expressions. Experimental results on both in-the-lab and in-the-wild compound expression datasets demonstrate the superiority of our proposed method against several state-of-the-art methods.