As one kind of architecture from the deep learning family, deep semantic segmentation network (DSSN) achieves a certain degree of success on the semantic segmentation task and obviously outperforms the traditional methods based on hand-crafted features. As a classic data-driven technique, DSSN can be trained by an end-to-end mechanism and competent for employing the low-level and mid-level cues (i.e., the discriminative image structure) to understand images, but lacks the high-level inference ability. By contrast, human beings have an excellent inference capacity and can be able to reliably interpret the RS imagery only when human beings master the basic RS domain knowledge. In literature, ontological modeling and reasoning is an ideal way to imitate and employ the domain knowledge of human beings, but is still rarely explored and adopted in the RS domain. To remedy the aforementioned critical limitation of DSSN, this paper proposes a collaboratively boosting framework (CBF) to combine data-driven deep learning module and knowledge-guided ontological reasoning module in an iterative way.

This paper proposes a reinforcement learning approach for nightly offline rebalancing operations in free-floating electric vehicle sharing systems (FFEVSS). Due to sparse demand in a network, FFEVSS require relocation of electrical vehicles (EVs) to charging stations and demander nodes, which is typically done by a group of drivers. A shuttle is used to pick up and drop off drivers throughout the network. The objective of this study is to solve the shuttle routing problem to finish the rebalancing work in the minimal time. We consider a reinforcement learning framework for the problem, in which a central controller determines the routing policies of a fleet of multiple shuttles. We deploy a policy gradient method for training recurrent neural networks and compare the obtained policy results with heuristic solutions. Our numerical studies show that unlike the existing solutions in the literature, the proposed methods allow to solve the general version of the problem with no restrictions on the urban EV network structure and charging requirements of EVs. Moreover, the learned policies offer a wide range of flexibility resulting in a significant reduction in the time needed to rebalance the network.

This paper frames a general prediction system as an observer traveling around a continuous space, measuring values at some locations, and predicting them at others. The observer is completely agnostic about any particular task being solved; it cares only about measurement locations and their values. This perspective leads to a machine learning framework in which seemingly unrelated tasks can be solved by a single model, by embedding their input and output variables into a shared space. An implementation of the framework is developed in which these variable embeddings are learned jointly with internal model parameters. In experiments, the approach is shown to (1) recover intuitive locations of variables in space and time, (2) exploit regularities across related datasets with completely disjoint input and output spaces, and (3) exploit regularities across seemingly unrelated tasks, outperforming task-specific single-task models and multi-task learning alternatives. The results suggest that even seemingly unrelated tasks may originate from similar underlying processes, a fact that the traveling observer model can use to make better predictions.

Molecule optimization is a fundamental task for accelerating drug discovery, with the goal of generating new valid molecules that maximize multiple drug properties while maintaining similarity to the input molecule. Existing generative models and reinforcement learning approaches made initial success, but still face difficulties in simultaneously optimizing multiple drug properties. To address such challenges, we propose the MultI-constraint MOlecule SAmpling (MIMOSA) approach, a sampling framework to use input molecule as an initial guess and sample molecules from the target distribution. MIMOSA first pretrains two property agnostic graph neural networks (GNNs) for molecule topology and substructure-type prediction, where a substructure can be either atom or single ring. For each iteration, MIMOSA uses the GNNs' prediction and employs three basic substructure operations (add, replace, delete) to generate new molecules and associated weights. The weights can encode multiple constraints including similarity and drug property constraints, upon which we select promising molecules for next iteration. MIMOSA enables flexible encoding of multiple property- and similarity-constraints and can efficiently generate new molecules that satisfy various property constraints and achieved up to 49.6% relative improvement over the best baseline in terms of success rate.

We consider problems of making sequences of decisions to accomplish tasks, interacting via the medium of language. These problems are often tackled with reinforcement learning approaches. We find that these models do not generalize well when applied to novel task domains. However, the large amount of computation necessary to adequately train and explore the search space of sequential decision making, under a reinforcement learning paradigm, precludes the inclusion of large contextualized language models, which might otherwise enable the desired generalization ability. We introduce a teacher-student imitation learning methodology and a means of converting a reinforcement learning model into a natural language understanding model. Together, these methodologies enable the introduction of contextualized language models into the sequential decision making problem space. We show that models can learn faster and generalize more, leveraging both the imitation learning and the reformulation. Our models exceed teacher performance on various held-out decision problems, by up to 7% on in-domain problems and 24% on out-of-domain problems.

Intelligent agents need to generalize from past experience to achieve goals in complex environments. World models facilitate such generalization and allow learning behaviors from imagined outcomes to increase sample-efficiency. While learning world models from image inputs has recently become feasible for some tasks, modeling Atari games accurately enough to derive successful behaviors has remained an open challenge for many years. We introduce DreamerV2, a reinforcement learning agent that learns behaviors purely from predictions in the compact latent space of a powerful world model. The world model uses discrete representations and is trained separately from the policy. DreamerV2 constitutes the first agent that achieves human-level performance on the Atari benchmark of 55 tasks by learning behaviors inside a separately trained world model. With the same computational budget and wall-clock time, DreamerV2 reaches 200M frames and exceeds the final performance of the top single-GPU agents IQN and Rainbow.

We show how feature maps in convolutional networks are susceptible to spatial bias. Due to a combination of architectural choices, the activation at certain locations is systematically elevated or weakened. The major source of this bias is the padding mechanism. Depending on several aspects of convolution arithmetic, this mechanism can apply the padding unevenly, leading to asymmetries in the learned weights. We demonstrate how such bias can be detrimental to certain tasks such as small object detection: the activation is suppressed if the stimulus lies in the impacted area, leading to blind spots and misdetection. We propose solutions to mitigate spatial bias and demonstrate how they can improve model accuracy. Convolutional neural networks (CNNs) have become state-of-the-art feature extractors for a wide variety of machine-learning tasks. A large body of work has focused on understanding the feature maps a CNN computes for an input. However, little attention has been paid to the spatial distribution of activation in the maps. Our interest in analyzing this distribution is triggered by mysterious failure cases of a traffic light detector: The detector is able to detect a small but visible traffic light with a high score in one frame of a road scene sequence. However, it fails completely in detecting the same traffic light in the next frame captured by the ego-vehicle.

The use of deep learning has grown at an exponential rate, giving rise to numerous specialized hardware and software systems for deep learning. Because the design space of deep learning software stacks and hardware accelerators is diverse and vast, prior work considers software optimizations separately from hardware architectures, effectively reducing the search space. Unfortunately, this bifurcated approach means that many profitable design points are never explored. This paper instead casts the problem as hardware/software co-design, with the goal of automatically identifying desirable points in the joint design space. The key to our solution is a new constrained Bayesian optimization framework that avoids invalid solutions by exploiting the highly constrained features of this design space, which are semicontinuous/semi-discrete. We evaluate our optimization framework by applying it to a variety of neural models, improving the energy-delay product by 18% (ResNet) and 40% (DQN) over hand-tuned state-of-the-art systems, as well as demonstrating strong results on other neural network architectures, such as MLPs and Transformers. The compute requirements of deep learning are growing at a double exponential rate (Hernandez & Brown, 2020), with more powerful models requiring exponentially more compute to train. This growth has been enabled by large systems of hardware accelerators, like GPUs and TPUs (NVIDIA, 2017; Jouppi et al., 2017). However, the continued scaling of these systems is limited by issues of power density, cooling, and memory, so we need to improve computational efficiency.

LECTRIC vehicles (EV) have been playing an increasingly important role in urban transportation and logistics tackle CO even without optimal labels. They consider solving systems for their capability of reducing greenhouse gas emission, problems through taking a sequence of actions similar to promoting renewable energy and introducing sustainable Markov decision process (MDP). Some reward schemes are transportation system [1], [2]. To model the operations of designed to inform the model about the quality of the actions logistic companies using EVs for service provision, Schneider it made based on which model parameters are adjusted to et al. proposed the electric vehicle routing problem with time enhance the solution quality. It has already been successfully windows (EVRPTW) [3]. In the context of EVRPTW, a fleet applied to various COs such as the travelling salesman problem of capacitated EVs is responsible for serving customers located (TSP), vehicle routing problem (VRP), minimum vertex cover in a specific region; each customer is associated with a demand (MVC), maximum cut (MAXCUT) etc. Despite the difficulty that must be satisfied during a time window; all the EVs are in training deep RL models, it is currently accepted as a very fully charged at the start of the planning horizon and could promising research direction to pursue.

Neural networks (NNs) whose subnetworks implement reusable functions are expected to offer numerous advantages, including compositionality through efficient recombination of functional building blocks, interpretability, preventing catastrophic interference, etc. Understanding if and how NNs are modular could provide insights into how to improve them. Current inspection methods, however, fail to link modules to their functionality. In this paper, we present a novel method based on learning binary weight masks to identify individual weights and subnets responsible for specific functions. Using this powerful tool, we contribute an extensive study of emerging modularity in NNs that covers several standard architectures and datasets. We demonstrate how common NNs fail to reuse submodules and offer new insights into the related issue of systematic generalization on language tasks.