Howie Choset is a Professor of Robotics where he serves as the co-director, along with Matt Travers, of the Biorobotics Lab. Choset's research program has made contributions to strategically significant problems in surgery, manufacturing, on-orbit maintenance, recycling and search and rescue. His work is most famous for its snake robots and other biologically inspired systems and recently his group has been contributing to robotic modularity, multi-agent planning, information-based search, and skill learning. Currently, Choset's projects include: medical support in the field, expeditionary robotics, on-orbit maintenance and construction of structures in space, rapidly carrying heavy objects up several flights of stairs, recycling of E-waste, food preparation, "edge"-sensing, and aerospace painting. Choset has led multi-PI projects centered on manufacturing: (1) automating the programming of robots for auto-body painting; (2) the development of mobile manipulators for agile and flexible fixture-free manufacturing of large structures in aerospace, and (3) the creation of a data-robot ecosystem for rapid manufacturing in the commercial electronics industry.

We then define abstract actions that correspond to control inputs that cause transitions between these regions. Due to the noise, every action has multiple possible outcomes that all occur with a certain probability. We compute lower and upper bounds (intervals) on these probabilities based on a finite number of observations of the noise. Our abstraction procedure ensures that we obtain a faithful, yet abstract representation of the autonomous system. In fact, this abstraction constitutes a type of Markov decision process, which is the standard type of model in sequential decision making under uncertainty. To analyze our abstract models in a rigorous manner, we use state-of-art tools from an area called formal verification.

This is the blog post version of the paper by the same name. Special thanks to Kate Woolverton, Paul Christiano, Rohin Shah, Alex Turner, William Saunders, Beth Barnes, Abram Demski, Scott Garrabrant, Sam Eisenstat, and Tsvi Benson-Tilsen for providing helpful comments and feedback on this post and the talk that preceded it. This post is a collection of 11 different proposals for building safe advanced AI under the current machine learning paradigm. There's a lot of literature out there laying out various different approaches such as amplification, debate, or recursive reward modeling, but a lot of that literature focuses primarily on outer alignment at the expense of inner alignment and doesn't provide direct comparisons between approaches. The goal of this post is to help solve that problem by providing a single collection of 11 different proposals for building safe advanced AI--each including both inner and outer alignment components. That being said, not only does this post not ...

In the Multi-Agent Path Finding (MAPF) problem, the goal is to find non-colliding paths for agents in an environment, such that each agent reaches its goal from its initial location. In safety-critical applications, a human supervisor may want to verify that the plan is indeed collision-free. To this end, a recent work introduces a notion of explainability for MAPF based on a visualization of the plan as a short sequence of images representing time segments, where in each time segment the trajectories of the agents are disjoint. Then, the explainable MAPF problem asks for a set of non-colliding paths that admits a short-enough explanation. Explainable MAPF adds a new difficulty to MAPF, in that it is NP-hard with respect to the size of the environment, and not just the number of agents. Thus, traditional MAPF algorithms are not equipped to directly handle explainable-MAPF. In this work, we adapt Conflict Based Search (CBS), a well-studied algorithm for MAPF, to handle explainable MAPF. We show how to add explainability constraints on top of the standard CBS tree and its underlying A* search. We examine the usefulness of this approach and, in particular, the tradeoff between planning time and explainability.

In this paper, we study a linear bandit optimization problem in a federated setting where a large collection of distributed agents collaboratively learn a common linear bandit model. Standard federated learning algorithms applied to this setting are vulnerable to Byzantine attacks on even a small fraction of agents. We propose a novel algorithm with a robust aggregation oracle that utilizes the geometric median. We prove that our proposed algorithm is robust to Byzantine attacks on fewer than half of agents and achieves a sublinear $\tilde{\mathcal{O}}({T^{3/4}})$ regret with $\mathcal{O}(\sqrt{T})$ steps of communication in $T$ steps. Moreover, we make our algorithm differentially private via a tree-based mechanism. Finally, if the level of corruption is known to be small, we show that using the geometric median of mean oracle for robust aggregation further improves the regret bound.

For example, consider the case where the perception module detects a pedestrian (PCV) on the road. It does not, however, recognize that the pedestrian is jaywalking. Even if no jaywalking events have been seen while training the CV perception module, representing knowledge of this event – i.e. (Pedestrian, participatesIn, Jaywalking) – in the scene KG could provide a new insight or cue for handling this edge-case with KEP (i.e.

Distributed learning paradigms, such as federated and decentralized learning, allow for the coordination of models across a collection of agents, and without the need to exchange raw data. Instead, agents compute model updates locally based on their available data, and subsequently share the update model with a parameter server or their peers. This is followed by an aggregation step, which traditionally takes the form of a (weighted) average. Distributed learning schemes based on averaging are known to be susceptible to outliers. A single malicious agent is able to drive an averaging-based distributed learning algorithm to an arbitrarily poor model. This has motivated the development of robust aggregation schemes, which are based on variations of the median and trimmed mean. While such procedures ensure robustness to outliers and malicious behavior, they come at the cost of significantly reduced sample efficiency. This means that current robust aggregation schemes require significantly higher agent participation rates to achieve a given level of performance than their mean-based counterparts in non-contaminated settings. In this work we remedy this drawback by developing statistically efficient and robust aggregation schemes for distributed learning.

Social choice functions (SCFs) map the preferences of a group of agents over some set of alternatives to a non-empty subset of alternatives. The Gibbard-Satterthwaite theorem has shown that only extremely restrictive SCFs are strategyproof when there are more than two alternatives. For set-valued SCFs, or so-called social choice correspondences, the situation is less clear. There are miscellaneous -- mostly negative -- results using a variety of strategyproofness notions and additional requirements. The simple and intuitive notion of Kelly-strategyproofness has turned out to be particularly compelling because it is weak enough to still allow for positive results. For example, the Pareto rule is strategyproof even when preferences are weak, and a number of attractive SCFs (such as the top cycle, the uncovered set, and the essential set) are strategyproof for strict preferences. In this paper, we show that, for weak preferences, only indecisive SCFs can satisfy strategyproofness. In particular, (i) every strategyproof rank-based SCF violates Pareto-optimality, (ii) every strategyproof support-based SCF (which generalize Fishburn's C2 SCFs) that satisfies Pareto-optimality returns at least one most preferred alternative of every voter, and (iii) every strategyproof non-imposing SCF returns the Condorcet loser in at least one profile. We also discuss the consequences of these results for randomized social choice.

Sender-receiver interactions, and specifically persuasion games, are widely researched in economic modeling and artificial intelligence, and serve as a solid foundation for powerful applications. However, in the classic persuasion games setting, the messages sent from the expert to the decision-maker are abstract or well-structured application-specific signals rather than natural (human) language messages, although natural language is a very common communication signal in real-world persuasion setups. This paper addresses the use of natural language in persuasion games, exploring its impact on the decisions made by the players and aiming to construct effective models for the prediction of these decisions. For this purpose, we conduct an online repeated interaction experiment. At each trial of the interaction, an informed expert aims to sell an uninformed decision-maker a vacation in a hotel, by sending her a review that describes the hotel. While the expert is exposed to several scored reviews, the decision-maker observes only the single review sent by the expert, and her payoff in case she chooses to take the hotel is a random draw from the review score distribution available to the expert only. The expert’s payoff, in turn, depends on the number of times the decision-maker chooses the hotel. We also compare the behavioral patterns in this experiment to the equivalent patterns in similar experiments where the communication is based on the numerical values of the reviews rather than the reviews’ text, and observe substantial differences which can be explained through an equilibrium analysis of the game. We consider a number of modeling approaches for our verbal communication setup, differing from each other in the model type (deep neural network (DNN) vs. linear classifier), the type of features used by the model (textual, behavioral or both) and the source of the textual features (DNN-based vs. hand-crafted). Our results demonstrate that given a prefix of the interaction sequence, our models can predict the future decisions of the decision-maker, particularly when a sequential modeling approach and hand-crafted textual features are applied. Further analysis of the hand-crafted textual features allows us to make initial observations about the aspects of text that drive decision making in our setup.

Here's a fun game: Name one person you know who enjoys sitting on hold, waiting for a customer service agent to resolve an issue or answer a question. AI can transform your customer service experience--if you implement it with care. But what if I were to say that AI can enable a world where you'll never again be held hostage on hold, wincing every time the smooth jazz version of Duran Duran's "Hungry Like the Wolf" repeats ad nauseum? It's true--customer service and support is a prime candidate for using AI to improve the customer experience and enable the team to scale. As the person responsible for operationalizing AI within ServiceNow's customer support experience, I seek out stages in the customer journey where AI can automate processes.