26 June 2023 in Proceedings of the AAAI Conference on Artificial Intelligence
The Abstraction and Reasoning Corpus (ARC) aims at benchmarking the performance of general artificial intelligence algorithms. The ARC's focus on broad generalization and few-shot learning has made it difficult to solve using pure machine learning. A more promising approach has been to perform program synthesis within an appropriately designed Domain Specific Language (DSL). However, these too have seen limited success. We propose Abstract Reasoning with Graph Abstractions (ARGA), a new object-centric framework that first represents images using graphs and then performs a search for a correct program in a DSL that is based on the abstracted graph space. The complexity of this combinatorial search is tamed through the use of constraint acquisition, state hashing, and Tabu search. An extensive set of experiments demonstrates the promise of ARGA in tackling some of the complicated object-centric tasks of the ARC rather efficiently, producing programs that are correct and easy to understand.
20 July 2020 in Journal of Artificial Intelligence Research
In many complex planning problems with factored, continuous state and action spaces such as Reservoir Control, Heating Ventilation and Air Conditioning (HVAC), and Navigation domains, it is difficult to obtain a model of the complex nonlinear dynamics that govern state evolution. However, the ubiquity of modern sensors allows us to collect large quantities of data from each of these complex systems and build accurate, nonlinear deep neural network models of their state transitions. But there remains one major problem for the task of control – how can we plan with deep network learned transition models without resorting to Monte Carlo Tree Search and other black-box transition model techniques that ignore model structure and do not easily extend to continuous domains? In this paper, we introduce two types of planning methods that can leverage deep neural network learned transition models: Hybrid Deep MILP Planner (HD-MILP-Plan) and Tensorflow Planner (TF-Plan). In HD-MILP-Plan, we make the critical observation that the Rectified Linear Unit (ReLU) transfer function for deep networks not only allows faster convergence of model learning, but also permits a direct compilation of the deep network transition model to a Mixed-Integer Linear Program (MILP) encoding. Further, we identify deep network specific optimizations for HD-MILP-Plan that improve performance over a base encoding and show that we can plan optimally with respect to the learned deep networks. In TF-Plan, we take advantage of the efficiency of auto-differentiation tools and GPU-based computation where we encode a subclass of purely continuous planning problems as Recurrent Neural Networks and directly optimize the actions through backpropagation. We compare both planners and show that TF-Plan is able to approximate the optimal plans found by HD-MILP-Plan in less computation time. Hence this article offers two novel planners for continuous state and action domains with learned deep neural net transition models: one optimal method (HD-MILP-Plan) and a scalable alternative for large-scale problems (TF-Plan).
1 June 2020 in Canadian Journal of Cardiology
17 September 2020 in Science and Technology for the Built Environment
Programmable thermostats with frequent schedule overrides (i.e., “holds”) are widely assumed to be detrimental to potential energy savings. We explored whether this assumption generalized to a population of smart thermostats using a longitudinal dataset of 20,000 devices from a single manufacturer. We observed a majority of devices were infrequently overridden, most often to less energy-efficient setpoints; nonetheless, the short duration of each override resulted in minimal increase in energy use. When we instead looked at the minority of thermostats consistently overridden (accounting for approximately 12% of the thermostats), we observed that a large fraction were being repeatedly manually adjusted, but often in an energy saving manner yielding energy performance on par with rarely overridden thermostats. Only the consistently overridden but infrequently adjusted population of thermostats showed poor energy performance, but these accounted for less than 5% of thermostats with energy performance roughly 2% worse on heating and 4% worse on cooling than the remaining population. While there may be effective ways to modify thermostat user interfaces to dissuade users from long-duration energy-inefficient hold settings, we conclude from our study of smart thermostats that user holding behavior may not be as detrimental to collective energy savings as commonly believed.