The purpose of this tutorial is to help individuals use the \underline{FireCommander} game environment for research applications. The FireCommander is an interactive, probabilistic joint perception-action reconnaissance environment in which a composite team of agents (e.g., robots) cooperate to fight dynamic, propagating firespots (e.g., targets). In FireCommander game, a team of agents must be tasked to optimally deal with a wildfire situation in an environment with propagating fire areas and some facilities such as houses, hospitals, power stations, etc. The team of agents can accomplish their mission by first sensing (e.g., estimating fire states), communicating the sensed fire-information among each other and then taking action to put the firespots out based on the sensed information (e.g., dropping water on estimated fire locations). The FireCommander environment can be useful for research topics spanning a wide range of applications from Reinforcement Learning (RL) and Learning from Demonstration (LfD), to Coordination, Psychology, Human-Robot Interaction (HRI) and Teaming. There are four important facets of the FireCommander environment that overall, create a non-trivial game: (1) Complex Objectives: Multi-objective Stochastic Environment, (2)Probabilistic Environment: Agents' actions result in probabilistic performance, (3) Hidden Targets: Partially Observable Environment and, (4) Uni-task Robots: Perception-only and Action-only agents. The FireCommander environment is first-of-its-kind in terms of including Perception-only and Action-only agents for coordination. It is a general multi-purpose game that can be useful in a variety of combinatorial optimization problems and stochastic games, such as applications of Reinforcement Learning (RL), Learning from Demonstration (LfD) and Inverse RL (iRL).

A new study confirms the success of a natural-gas leak-detection tool pioneered by Los Alamos National Laboratory scientists that uses sensors and machine learning to locate leak points at oil and gas fields, promising new automatic, affordable sampling across vast natural gas infrastructure. "Our automated leak location system finds gas leaks fast, including small ones from failing infrastructure, and lowers cost as current methods to fix gas leaks are labor intensive, expensive and slow," said Manvendra Dubey, the lead Los Alamos National Laboratory scientist and coauthor of the new study. "Our sensors outperformed competing techniques in sensitivity to detecting methane and ethane. In addition, our neural network can be coupled to any sensor, which makes our tool very powerful and will enable market penetration." The Autonomous, Low-cost, Fast Leak Detection System (ALFaLDS) was developed to discover accidental releases of methane, a potent greenhouse gas, and won a 2019 R&D 100 award.

October 28, 2020 – "It has been a wild year," says Montreal-based Brainbox AI's president, Sam Ramadori. When the company launched in May 2019, it had just shy of 20 people, and worked on six development buildings totalling half a million square feet. "Today, we're in over 30 million, and with the commitment here, getting us over 40 million installed square feet." The commitment to which Ramadori is referring is that of AMP Capital in Australia, which is set to roll out Brainbox AI's energy-saving AI technology--which promises a reduction of carbon footprint by 20-40% and 15-20% in energy spend--across its entire managed real estate portfolio of central business district (CBD) office buildings, retail shopping centres and logistics facilities. Six AMP Capital-managed buildings across Australia are already trialling the technology, which will then be rolled out across the entire portfolio of over 40 buildings in Australia and New Zealand by Q1 2021. "We have a very young team, of very highly-capable technical folks, and they're all driven, not only by the truly autonomous solution, but we can really make a strong impact on climate change… but the only way to do that is to be able to scale quickly," said Ramadori, referring to the solution's ease of deployment.

We study the control of an \emph{unknown} linear dynamical system under general convex costs. The objective is minimizing regret vs. the class of disturbance-feedback-controllers, which encompasses all stabilizing linear-dynamical-controllers. In this work, we first consider the case of known cost functions, for which we design the first polynomial-time algorithm with $n^3\sqrt{T}$-regret, where $n$ is the dimension of the state plus the dimension of control input. The $\sqrt{T}$-horizon dependence is optimal, and improves upon the previous best known bound of $T^{2/3}$. The main component of our algorithm is a novel geometric exploration strategy: we adaptively construct a sequence of barycentric spanners in the policy space. Second, we consider the case of bandit feedback, for which we give the first polynomial-time algorithm with $poly(n)\sqrt{T}$-regret, building on Stochastic Bandit Convex Optimization.

Gaussian processes are a versatile framework for learning unknown functions in a manner that permits one to utilize prior information about their properties. Although many different Gaussian process models are readily available when the input space is Euclidean, the choice is much more limited for Gaussian processes whose input space is an undirected graph. In this work, we leverage the stochastic partial differential equation characterization of Mat\'{e}rn Gaussian processes - a widely-used model class in the Euclidean setting - to study their analog for undirected graphs. We show that the resulting Gaussian processes inherit various attractive properties of their Euclidean and Riemannian analogs and provide techniques that allow them to be trained using standard methods, such as inducing points. This enables graph Mat\'{e}rn Gaussian processes to be employed in mini-batch and non-conjugate settings, thereby making them more accessible to practitioners and easier to deploy within larger learning frameworks.

We study a multi-agent stochastic linear bandit with side information, parameterized by an unknown vector $\theta^* \in \mathbb{R}^d$. The side information consists of a finite collection of low-dimensional subspaces, one of which contains $\theta^*$. In our setting, agents can collaborate to reduce regret by sending recommendations across a communication graph connecting them. We present a novel decentralized algorithm, where agents communicate subspace indices with each other, and each agent plays a projected variant of LinUCB on the corresponding (low-dimensional) subspace. Through a combination of collaborative best subspace identification, and per-agent learning of an unknown vector in the corresponding low-dimensional subspace, we show that the per-agent regret is much smaller than the case when agents do not communicate. By collaborating to identify the subspace containing $\theta^*$, we show that each agent effectively solves an easier instance of the linear bandit (compared to the case of no collaboration), thus leading to the reduced per-agent regret. We finally complement these results through simulations.

Model selection has been proven an effective strategy for improving accuracy in time series forecasting applications. However, when dealing with hierarchical time series, apart from selecting the most appropriate forecasting model, forecasters have also to select a suitable method for reconciling the base forecasts produced for each series to make sure they are coherent. Although some hierarchical forecasting methods like minimum trace are strongly supported both theoretically and empirically for reconciling the base forecasts, there are still circumstances under which they might not produce the most accurate results, being outperformed by other methods. In this paper we propose an approach for dynamically selecting the most appropriate hierarchical forecasting method and succeeding better forecasting accuracy along with coherence. The approach, to be called conditional hierarchical forecasting, is based on Machine Learning classification methods and uses time series features as leading indicators for performing the selection for each hierarchy examined considering a variety of alternatives. Our results suggest that conditional hierarchical forecasting leads to significantly more accurate forecasts than standard approaches, especially at lower hierarchical levels.

Exponential families are widely used in machine learning; they include many distributions in continuous and discrete domains (e.g., Gaussian, Dirichlet, Poisson, and categorical distributions via the softmax transformation). Distributions in each of these families have fixed support. In contrast, for finite domains, there has been recent work on sparse alternatives to softmax (e.g. sparsemax and alpha-entmax), which have varying support, being able to assign zero probability to irrelevant categories. This paper expands that work in two directions: first, we extend alpha-entmax to continuous domains, revealing a link with Tsallis statistics and deformed exponential families. Second, we introduce continuous-domain attention mechanisms, deriving efficient gradient backpropagation algorithms for alpha in {1,2}. Experiments on attention-based text classification, machine translation, and visual question answering illustrate the use of continuous attention in 1D and 2D, showing that it allows attending to time intervals and compact regions.

Spiking neural networks (SNNs) are promising in a bio-plausible coding for spatio-temporal information and event-driven signal processing, which is very suited for energy-efficient implementation in neuromorphic hardware. However, the unique working mode of SNNs makes them more difficult to train than traditional networks. Currently, there are two main routes to explore the training of deep SNNs with high performance. The first is to convert a pre-trained ANN model to its SNN version, which usually requires a long coding window for convergence and cannot exploit the spatio-temporal features during training for solving temporal tasks. The other is to directly train SNNs in the spatio-temporal domain. But due to the binary spike activity of the firing function and the problem of gradient vanishing or explosion, current methods are restricted to shallow architectures and thereby difficult in harnessing large-scale datasets (e.g. ImageNet). To this end, we propose a threshold-dependent batch normalization (tdBN) method based on the emerging spatio-temporal backpropagation, termed "STBP-tdBN", enabling direct training of a very deep SNN and the efficient implementation of its inference on neuromorphic hardware. With the proposed method and elaborated shortcut connection, we significantly extend directly-trained SNNs from a shallow structure ( < 10 layer) to a very deep structure (50 layers). Furthermore, we theoretically analyze the effectiveness of our method based on "Block Dynamical Isometry" theory. Finally, we report superior accuracy results including 93.15 % on CIFAR-10, 67.8 % on DVS-CIFAR10, and 67.05% on ImageNet with very few timesteps. To our best knowledge, it's the first time to explore the directly-trained deep SNNs with high performance on ImageNet.

The Admissions team will respond to inquiries sent to this email address within two business days (Monday through Friday from 8 a.m. to 5 p.m. Pacific Time). Select the info session you would like to attend (Required): Select the info session you would like to attend (Required): November 9 at 5 p.m. PST How did you hear about the Master of Engineering (MEng) program? How did you hear about the Master of Engineering (MEng) program? Questions you would like to ask in advance of the info session?