Siddharth Reddy

Siddharth Reddy, Sergey Levine, Anca D. Dragan

How can we train an assistive human-machine interface (e.g., an electromyography-based limb prosthesis) to translate a user's raw command signals into the actions of a robot or computer when there is no prior mapping, we cannot ask the user for supervision in the form of action labels or reward feedback, and we do not have prior knowledge of the tasks the user is trying to accomplish? The key idea in this paper is that, regardless of the task, when an interface is more intuitive, the user's commands are less noisy. We formalize this idea as a completely unsupervised objective for optimizing interfaces: the mutual information between the user's command signals and the induced state transitions in the environment. To evaluate whether this mutual information score can distinguish between effective and ineffective interfaces, we conduct an observational study on 540K examples of users operating various keyboard and eye gaze interfaces for typing, controlling simulated robots, and playing video games. The results show that our mutual information scores are predictive of the ground-truth task completion metrics in a variety of domains, with an average Spearman's rank correlation of 0.43. In addition to offline evaluation of existing interfaces, we use our unsupervised objective to learn an interface from scratch: we randomly initialize the interface, have the user attempt to perform their desired tasks using the interface, measure the mutual information score, and update the interface to maximize mutual information through reinforcement learning. We evaluate our method through a user study with 12 participants who perform a 2D cursor control task using a perturbed mouse, and an experiment with one user playing the Lunar Lander game using hand gestures. The results show that we can learn an interface from scratch, without any user supervision or prior knowledge of tasks, in under 30 minutes.

Jensen Gao, Siddharth Reddy, Glen Berseth et al.

We aim to help users communicate their intent to machines using flexible, adaptive interfaces that translate arbitrary user input into desired actions. In this work, we focus on assistive typing applications in which a user cannot operate a keyboard, but can instead supply other inputs, such as webcam images that capture eye gaze or neural activity measured by a brain implant. Standard methods train a model on a fixed dataset of user inputs, then deploy a static interface that does not learn from its mistakes; in part, because extracting an error signal from user behavior can be challenging. We investigate a simple idea that would enable such interfaces to improve over time, with minimal additional effort from the user: online learning from user feedback on the accuracy of the interface's actions. In the typing domain, we leverage backspaces as feedback that the interface did not perform the desired action. We propose an algorithm called x-to-text (X2T) that trains a predictive model of this feedback signal, and uses this model to fine-tune any existing, default interface for translating user input into actions that select words or characters. We evaluate X2T through a small-scale online user study with 12 participants who type sentences by gazing at their desired words, a large-scale observational study on handwriting samples from 60 users, and a pilot study with one participant using an electrocorticography-based brain-computer interface. The results show that X2T learns to outperform a non-adaptive default interface, stimulates user co-adaptation to the interface, personalizes the interface to individual users, and can leverage offline data collected from the default interface to improve its initial performance and accelerate online learning.

Jensen Gao, Siddharth Reddy, Glen Berseth et al.

Adaptive interfaces can help users perform sequential decision-making tasks like robotic teleoperation given noisy, high-dimensional command signals (e.g., from a brain-computer interface). Recent advances in human-in-the-loop machine learning enable such systems to improve by interacting with users, but tend to be limited by the amount of data that they can collect from individual users in practice. In this paper, we propose a reinforcement learning algorithm to address this by training an interface to map raw command signals to actions using a combination of offline pre-training and online fine-tuning. To address the challenges posed by noisy command signals and sparse rewards, we develop a novel method for representing and inferring the user's long-term intent for a given trajectory. We primarily evaluate our method's ability to assist users who can only communicate through noisy, high-dimensional input channels through a user study in which 12 participants performed a simulated navigation task by using their eye gaze to modulate a 128-dimensional command signal from their webcam. The results show that our method enables successful goal navigation more often than a baseline directional interface, by learning to denoise user commands signals and provide shared autonomy assistance. We further evaluate on a simulated Sawyer pushing task with eye gaze control, and the Lunar Lander game with simulated user commands, and find that our method improves over baseline interfaces in these domains as well. Extensive ablation experiments with simulated user commands empirically motivate each component of our method.

Sean Chen, Jensen Gao, Siddharth Reddy et al.

Building assistive interfaces for controlling robots through arbitrary, high-dimensional, noisy inputs (e.g., webcam images of eye gaze) can be challenging, especially when it involves inferring the user's desired action in the absence of a natural 'default' interface. Reinforcement learning from online user feedback on the system's performance presents a natural solution to this problem, and enables the interface to adapt to individual users. However, this approach tends to require a large amount of human-in-the-loop training data, especially when feedback is sparse. We propose a hierarchical solution that learns efficiently from sparse user feedback: we use offline pre-training to acquire a latent embedding space of useful, high-level robot behaviors, which, in turn, enables the system to focus on using online user feedback to learn a mapping from user inputs to desired high-level behaviors. The key insight is that access to a pre-trained policy enables the system to learn more from sparse rewards than a naïve RL algorithm: using the pre-trained policy, the system can make use of successful task executions to relabel, in hindsight, what the user actually meant to do during unsuccessful executions. We evaluate our method primarily through a user study with 12 participants who perform tasks in three simulated robotic manipulation domains using a webcam and their eye gaze: flipping light switches, opening a shelf door to reach objects inside, and rotating a valve. The results show that our method successfully learns to map 128-dimensional gaze features to 7-dimensional joint torques from sparse rewards in under 10 minutes of online training, and seamlessly helps users who employ different gaze strategies, while adapting to distributional shift in webcam inputs, tasks, and environments.

Siddharth Reddy, Anca D. Dragan, Sergey Levine

Standard lossy image compression algorithms aim to preserve an image's appearance, while minimizing the number of bits needed to transmit it. However, the amount of information actually needed by a user for downstream tasks -- e.g., deciding which product to click on in a shopping website -- is likely much lower. To achieve this lower bitrate, we would ideally only transmit the visual features that drive user behavior, while discarding details irrelevant to the user's decisions. We approach this problem by training a compression model through human-in-the-loop learning as the user performs tasks with the compressed images. The key insight is to train the model to produce a compressed image that induces the user to take the same action that they would have taken had they seen the original image. To approximate the loss function for this model, we train a discriminator that tries to distinguish whether a user's action was taken in response to the compressed image or the original. We evaluate our method through experiments with human participants on four tasks: reading handwritten digits, verifying photos of faces, browsing an online shopping catalogue, and playing a car racing video game. The results show that our method learns to match the user's actions with and without compression at lower bitrates than baseline methods, and adapts the compression model to the user's behavior: it preserves the digit number and randomizes handwriting style in the digit reading task, preserves hats and eyeglasses while randomizing faces in the photo verification task, preserves the perceived price of an item while randomizing its color and background in the online shopping task, and preserves upcoming bends in the road in the car racing game.

Siddharth Reddy, Sergey Levine, Anca D. Dragan

We aim to help users estimate the state of the world in tasks like robotic teleoperation and navigation with visual impairments, where users may have systematic biases that lead to suboptimal behavior: they might struggle to process observations from multiple sensors simultaneously, receive delayed observations, or overestimate distances to obstacles. While we cannot directly change the user's internal beliefs or their internal state estimation process, our insight is that we can still assist them by modifying the user's observations. Instead of showing the user their true observations, we synthesize new observations that lead to more accurate internal state estimates when processed by the user. We refer to this method as assistive state estimation (ASE): an automated assistant uses the true observations to infer the state of the world, then generates a modified observation for the user to consume (e.g., through an augmented reality interface), and optimizes the modification to induce the user's new beliefs to match the assistant's current beliefs. We evaluate ASE in a user study with 12 participants who each perform four tasks: two tasks with known user biases -- bandwidth-limited image classification and a driving video game with observation delay -- and two with unknown biases that our method has to learn -- guided 2D navigation and a lunar lander teleoperation video game. A different assistance strategy emerges in each domain, such as quickly revealing informative pixels to speed up image classification, using a dynamics model to undo observation delay in driving, identifying nearby landmarks for navigation, and exaggerating a visual indicator of tilt in the lander game. The results show that ASE substantially improves the task performance of users with bandwidth constraints, observation delay, and other unknown biases.

Siddharth Reddy, Anca D. Dragan, Sergey Levine et al.

We seek to align agent behavior with a user's objectives in a reinforcement learning setting with unknown dynamics, an unknown reward function, and unknown unsafe states. The user knows the rewards and unsafe states, but querying the user is expensive. To address this challenge, we propose an algorithm that safely and interactively learns a model of the user's reward function. We start with a generative model of initial states and a forward dynamics model trained on off-policy data. Our method uses these models to synthesize hypothetical behaviors, asks the user to label the behaviors with rewards, and trains a neural network to predict the rewards. The key idea is to actively synthesize the hypothetical behaviors from scratch by maximizing tractable proxies for the value of information, without interacting with the environment. We call this method reward query synthesis via trajectory optimization (ReQueST). We evaluate ReQueST with simulated users on a state-based 2D navigation task and the image-based Car Racing video game. The results show that ReQueST significantly outperforms prior methods in learning reward models that transfer to new environments with different initial state distributions. Moreover, ReQueST safely trains the reward model to detect unsafe states, and corrects reward hacking before deploying the agent.

Gokul Swamy, Siddharth Reddy, Sergey Levine et al.

Autonomous robots often encounter challenging situations where their control policies fail and an expert human operator must briefly intervene, e.g., through teleoperation. In settings where multiple robots act in separate environments, a single human operator can manage a fleet of robots by identifying and teleoperating one robot at any given time. The key challenge is that users have limited attention: as the number of robots increases, users lose the ability to decide which robot requires teleoperation the most. Our goal is to automate this decision, thereby enabling users to supervise more robots than their attention would normally allow for. Our insight is that we can model the user's choice of which robot to control as an approximately optimal decision that maximizes the user's utility function. We learn a model of the user's preferences from observations of the user's choices in easy settings with a few robots, and use it in challenging settings with more robots to automatically identify which robot the user would most likely choose to control, if they were able to evaluate the states of all robots at all times. We run simulation experiments and a user study with twelve participants that show our method can be used to assist users in performing a simulated navigation task. We also run a hardware demonstration that illustrates how our method can be applied to a real-world mobile robot navigation task.

Siddharth Reddy, Anca D. Dragan, Sergey Levine

Learning to imitate expert behavior from demonstrations can be challenging, especially in environments with high-dimensional, continuous observations and unknown dynamics. Supervised learning methods based on behavioral cloning (BC) suffer from distribution shift: because the agent greedily imitates demonstrated actions, it can drift away from demonstrated states due to error accumulation. Recent methods based on reinforcement learning (RL), such as inverse RL and generative adversarial imitation learning (GAIL), overcome this issue by training an RL agent to match the demonstrations over a long horizon. Since the true reward function for the task is unknown, these methods learn a reward function from the demonstrations, often using complex and brittle approximation techniques that involve adversarial training. We propose a simple alternative that still uses RL, but does not require learning a reward function. The key idea is to provide the agent with an incentive to match the demonstrations over a long horizon, by encouraging it to return to demonstrated states upon encountering new, out-of-distribution states. We accomplish this by giving the agent a constant reward of r=+1 for matching the demonstrated action in a demonstrated state, and a constant reward of r=0 for all other behavior. Our method, which we call soft Q imitation learning (SQIL), can be implemented with a handful of minor modifications to any standard Q-learning or off-policy actor-critic algorithm. Theoretically, we show that SQIL can be interpreted as a regularized variant of BC that uses a sparsity prior to encourage long-horizon imitation. Empirically, we show that SQIL outperforms BC and achieves competitive results compared to GAIL, on a variety of image-based and low-dimensional tasks in Box2D, Atari, and MuJoCo.

Siddharth Reddy, Anca D. Dragan, Sergey Levine

Inferring intent from observed behavior has been studied extensively within the frameworks of Bayesian inverse planning and inverse reinforcement learning. These methods infer a goal or reward function that best explains the actions of the observed agent, typically a human demonstrator. Another agent can use this inferred intent to predict, imitate, or assist the human user. However, a central assumption in inverse reinforcement learning is that the demonstrator is close to optimal. While models of suboptimal behavior exist, they typically assume that suboptimal actions are the result of some type of random noise or a known cognitive bias, like temporal inconsistency. In this paper, we take an alternative approach, and model suboptimal behavior as the result of internal model misspecification: the reason that user actions might deviate from near-optimal actions is that the user has an incorrect set of beliefs about the rules -- the dynamics -- governing how actions affect the environment. Our insight is that while demonstrated actions may be suboptimal in the real world, they may actually be near-optimal with respect to the user's internal model of the dynamics. By estimating these internal beliefs from observed behavior, we arrive at a new method for inferring intent. We demonstrate in simulation and in a user study with 12 participants that this approach enables us to more accurately model human intent, and can be used in a variety of applications, including offering assistance in a shared autonomy framework and inferring human preferences.

Siddharth Reddy, Anca D. Dragan, Sergey Levine

In shared autonomy, user input is combined with semi-autonomous control to achieve a common goal. The goal is often unknown ex-ante, so prior work enables agents to infer the goal from user input and assist with the task. Such methods tend to assume some combination of knowledge of the dynamics of the environment, the user's policy given their goal, and the set of possible goals the user might target, which limits their application to real-world scenarios. We propose a deep reinforcement learning framework for model-free shared autonomy that lifts these assumptions. We use human-in-the-loop reinforcement learning with neural network function approximation to learn an end-to-end mapping from environmental observation and user input to agent action values, with task reward as the only form of supervision. This approach poses the challenge of following user commands closely enough to provide the user with real-time action feedback and thereby ensure high-quality user input, but also deviating from the user's actions when they are suboptimal. We balance these two needs by discarding actions whose values fall below some threshold, then selecting the remaining action closest to the user's input. Controlled studies with users (n = 12) and synthetic pilots playing a video game, and a pilot study with users (n = 4) flying a real quadrotor, demonstrate the ability of our algorithm to assist users with real-time control tasks in which the agent cannot directly access the user's private information through observations, but receives a reward signal and user input that both depend on the user's intent. The agent learns to assist the user without access to this private information, implicitly inferring it from the user's input. This paper is a proof of concept that illustrates the potential for deep reinforcement learning to enable flexible and practical assistive systems.

Siddharth Reddy, Igor Labutov, Siddhartha Banerjee et al.

In the study of human learning, there is broad evidence that our ability to retain information improves with repeated exposure and decays with delay since last exposure. This plays a crucial role in the design of educational software, leading to a trade-off between teaching new material and reviewing what has already been taught. A common way to balance this trade-off is spaced repetition, which uses periodic review of content to improve long-term retention. Though spaced repetition is widely used in practice, e.g., in electronic flashcard software, there is little formal understanding of the design of these systems. Our paper addresses this gap in three ways. First, we mine log data from spaced repetition software to establish the functional dependence of retention on reinforcement and delay. Second, we use this memory model to develop a stochastic model for spaced repetition systems. We propose a queueing network model of the Leitner system for reviewing flashcards, along with a heuristic approximation that admits a tractable optimization problem for review scheduling. Finally, we empirically evaluate our queueing model through a Mechanical Turk experiment, verifying a key qualitative prediction of our model: the existence of a sharp phase transition in learning outcomes upon increasing the rate of new item introductions.

Siddharth Reddy, Igor Labutov, Thorsten Joachims

Students in online courses generate large amounts of data that can be used to personalize the learning process and improve quality of education. In this paper, we present the Latent Skill Embedding (LSE), a probabilistic model of students and educational content that can be used to recommend personalized sequences of lessons with the goal of helping students prepare for specific assessments. Akin to collaborative filtering for recommender systems, the algorithm does not require students or content to be described by features, but it learns a representation using access traces. We formulate this problem as a regularized maximum-likelihood embedding of students, lessons, and assessments from historical student-content interactions. An empirical evaluation on large-scale data from Knewton, an adaptive learning technology company, shows that this approach predicts assessment results competitively with benchmark models and is able to discriminate between lesson sequences that lead to mastery and failure.