In this paper, we introduce a distributed version of the classical stochastic Multi-Arm Bandit (MAB) problem. Our setting consists of a large number of agents $n$ that collaboratively and simultaneously solve the same instance of $K$ armed MAB to minimize the average cumulative regret over all agents. The agents can communicate and collaborate among each other \emph{only} through a pairwise asynchronous gossip based protocol that exchange a limited number of bits. In our model, agents at each point decide on (i) which arm to play, (ii) whether to, and if so (iii) what and whom to communicate with. Agents in our model are decentralized, namely their actions only depend on their observed history in the past. We develop a novel algorithm in which agents, whenever they choose, communicate only arm-ids and not samples, with another agent chosen uniformly and independently at random. The per-agent regret scaling achieved by our algorithm is $O \left( \frac{\lceil\frac{K}{n}\rceil+\log(n)}{\Delta} \log(T) + \frac{\log^3(n) \log \log(n)}{\Delta^2} \right)$. Furthermore, any agent in our algorithm communicates only a total of $\Theta(\log(T))$ times over a time interval of $T$. We compare our results to two benchmarks - one where there is no communication among agents and one corresponding to complete interaction. We show both theoretically and empirically, that our algorithm experiences a significant reduction both in per-agent regret when compared to the case when agents do not collaborate and in communication complexity when compared to the full interaction setting which requires $T$ communication attempts by an agent over $T$ arm pulls.

In this paper, we study the problem of computing $U$-statistics of degree $2$, i.e., quantities that come in the form of averages over pairs of data points, in the local model of differential privacy (LDP). The class of $U$-statistics covers many statistical estimates of interest, including Gini mean difference, Kendall's tau coefficient and Area under the ROC Curve (AUC), as well as empirical risk measures for machine learning problems such as ranking, clustering and metric learning. We first introduce an LDP protocol based on quantizing the data into bins and applying randomized response, which guarantees an $\epsilon$-LDP estimate with a Mean Squared Error (MSE) of $O(1/\sqrt{n}\epsilon)$ under regularity assumptions on the $U$-statistic or the data distribution. We then propose a specialized protocol for AUC based on a novel use of hierarchical histograms that achieves MSE of $O(\alpha^3/n\epsilon^2)$ for arbitrary data distribution. We also show that 2-party secure computation allows to design a protocol with MSE of $O(1/n\epsilon^2)$, without any assumption on the kernel function or data distribution and with total communication linear in the number of users $n$. Finally, we evaluate the performance of our protocols through experiments on synthetic and real datasets.

Federated learning (FL) provides a communication-efficient approach to solve machine learning problems concerning distributed data, without sending raw data to a central server. However, existing works on FL only utilize first-order gradient descent (GD) and do not consider the preceding iterations to gradient update which can potentially accelerate convergence. In this paper, we consider momentum term which relates to the last iteration. The proposed momentum federated learning (MFL) uses momentum gradient descent (MGD) in the local update step of FL system. We establish global convergence properties of MFL and derive an upper bound on MFL convergence rate. Comparing the upper bounds on MFL and FL convergence rate, we provide conditions in which MFL accelerates the convergence. For different machine learning models, the convergence performance of MFL is evaluated based on experiments with MNIST dataset. Simulation results comfirm that MFL is globally convergent and further reveal significant convergence improvement over FL.

A BSTRACT Biological evolution has distilled the experiences of many learners into the general learning algorithms of humans. Our novel meta-reinforcement learning algorithm MetaGenRL is inspired by this process. MetaGenRL distills the experiences of many complex agents to meta-learn a low-complexity neural objective function that affects how future individuals will learn. Unlike recent meta-RL algorithms, MetaGenRL can generalize to new environments that are entirely different from those used for meta-training. In some cases, it even outperforms human-engineered RL algorithms. MetaGenRL uses off-policy second-order gradients during meta-training that greatly increase its sample efficiency. 1 I NTRODUCTION The process of evolution has equipped humans with incredibly general learning algorithms. They allow us to flexibly solve a wide range of problems, even in the absence of many related prior experiences. The inductive biases that give rise to these capabilities are the result of distilling the collective experiences of many learners throughout the course of natural evolution. By essentially learning from learning experiences in this way, this knowledge can be compactly encoded in the genetic code of an individual to give rise to the general learning capabilities that we observe today. In contrast, Reinforcement Learning (RL) in artificial agents rarely proceeds in this way. The learning rules that are used to train agents are the result of years of human engineering and design, (e.g. Correspondingly, artificial agents are inherently limited by the ability of the designer to incorporate the right inductive biases in order to learn from previous experiences.

Over its lifetime, a reinforcement learning agent is often tasked with different tasks. How to efficiently adapt a previously learned control policy from one task to another, remains an open research question. In this paper, we investigate how instructions formulated in natural language can enable faster and more effective task adaptation. This can serve as the basis for developing language instructed skills, which can be used in a lifelong learning setting. Our method is capable of assessing, given a set of developed base control policies, which policy will adapt best to a new unseen task.

Evidence-based reasoning is at the core of ma ny problem - solving and decision-making tasks in a wide variety of domains. Generalizing from the research and development of cognitive agents in several such domains, this paper presents progress toward a computational theory for the development of instructable cognitive agents for evide nce-based reasoning tasks. The paper also illustrates the application of this theory to the development of four prototype cognitive agents in domains that are critical to the government and the public sector . Two agents function as cognitive assistants, one in intelligence analysis, and the other in science education . The other two agents operate autonomously, one in cybersecurity and the other in intelligence, surveillance, and reconnaissance. The paper concludes with the directions of future research on th e proposed computational theory.

Today, there's a common notion that artificial intelligence (AI) is going to put us all out of jobs. But recently, I read an article (subscription required) where AI expert Robert Atkinson remarked that worries around AI and job loss are overblown. He said, "It's time to take a deep breath and stop panicking about artificial intelligence and what it portends for jobs. No, AI won't destroy more jobs than it creates. No, the pace of technological change is not accelerating. And no, we certainly don't need to tax AI to slow it down."

Founded in 1990, the Martin Luther King Jr. (MLK) Visiting Professors and Scholars Program honors the life and legacy of Martin Luther King by increasing the presence of, and recognizing the contributions of, underrepresented minority scholars at MIT. MLK Visiting Professors and Scholars enhance their scholarship through intellectual engagement with the MIT community and enrich the cultural, academic, and professional experience of students. Six scholars are visiting MIT this academic year as part of the program. Kasso Okoudjou is returning for a second year as an MLK Visiting Professor in the Department of Mathematics. Originally from Benin, he moved to the United States in 1998 and earned a PhD in mathematics from Georgia Tech. Okoudjou joins MIT from the University of Maryland College Park, where he is a professor.

Earlier this year, President Trump signed an executive order for the "American AI Initiative," to guide AI developments and investments in the following areas: research and development, ethical standards, automation, and international outreach. This initiative is indicative of the changing times, and how, as a country, the U.S. is learning to navigate the implications of AI. Leaders in the business world, specifically, are faced with the responsibility of equipping our employees with the skills necessary for paving long-lasting career paths, and the workforce must discover what will be expected as technology continues to disrupt the norm, and work as we know it. As a global business leader, an AI optimist, and a father, I find myself asking: What will make a career sustainable in 2020 and beyond? Will the future of education rise to meet the demands of the future of work?