In order to recognize abnormal events such as traffic accidents in a video sequence, it is important that the network takes into account interactions of objects in a given frame. We propose a novel Feature Aggregation (FA) block that refines each object's features by computing a weighted sum of the features of all objects in a frame. We use FA block along with Long Short Term Memory (LSTM) network to anticipate accidents in the video sequences. We report mean Average Precision (mAP) and Average Time-to-Accident (ATTA) on Street Accident (SA) dataset. Our proposed method achieves the highest score for risk anticipation by predicting accidents 0.32 sec and 0.75 sec earlier compared to the best results with Adaptive Loss and dynamic parameter prediction based methods respectively.

Lio is a mobile robot platform with a multi-functional arm explicitly designed for human-robot interaction and personal care assistant tasks. The robot has already been deployed in several health care facilities, where it is functioning autonomously, assisting staff and patients on an everyday basis. Lio is intrinsically safe by having full coverage in soft artificial-leather material as well as having collision detection, limited speed and forces. Furthermore, the robot has a compliant motion controller. A combination of visual, audio, laser, ultrasound and mechanical sensors are used for safe navigation and environment understanding. The ROS-enabled setup allows researchers to access raw sensor data as well as have direct control of the robot. The friendly appearance of Lio has resulted in the robot being well accepted by health care staff and patients. Fully autonomous operation is made possible by a flexible decision engine, autonomous navigation and automatic recharging. Combined with time-scheduled task triggers, this allows Lio to operate throughout the day, with a battery life of up to 8 hours and recharging during idle times. A combination of powerful on-board computing units provides enough processing power to deploy artificial intelligence and deep learning-based solutions on-board the robot without the need to send any sensitive data to cloud services, guaranteeing compliance with privacy requirements. During the COVID-19 pandemic, Lio was rapidly adjusted to perform additional functionality like disinfection and remote elevated body temperature detection. It complies with ISO13482 - Safety requirements for personal care robots, meaning it can be directly tested and deployed in care facilities.

Fine-grained aspect extraction is an essential sub-task in aspect based opinion analysis. It aims to identify the aspect terms (a.k.a. opinion targets) of a product or service in each sentence. However, expensive annotation process is usually involved to acquire sufficient token-level labels for each domain. To address this limitation, some previous works propose domain adaptation strategies to transfer knowledge from a sufficiently labeled source domain to unlabeled target domains. But due to both the difficulty of fine-grained prediction problems and the large domain gap between domains, the performance remains unsatisfactory. This work conducts a pioneer study on leveraging sentence-level aspect category labels that can be usually available in commercial services like review sites to promote token-level transfer for the extraction purpose. Specifically, the aspect category information is used to construct pivot knowledge for transfer with assumption that the interactions between sentence-level aspect category and token-level aspect terms are invariant across domains. To this end, we propose a novel multi-level reconstruction mechanism that aligns both the fine-grained and coarse-grained information in multiple levels of abstractions. Comprehensive experiments demonstrate that our approach can fully utilize sentence-level aspect category labels to improve cross-domain aspect extraction with a large performance gain.

Neural Architecture Search (NAS) has gained attraction due to superior classification performance. Differential Architecture Search (DARTS) is a computationally light method. To limit computational resources DARTS makes numerous approximations. These approximations result in inferior performance. We propose to fine-tune DARTS using fixed operations as they are independent of these approximations. Our method offers a good trade-off between the number of parameters and classification accuracy. Our approach improves the top-1 accuracy on Fashion-MNIST, CompCars, and MIO-TCD datasets by 0.56%, 0.50%, and 0.39%, respectively compared to the state-of-the-art approaches. Our approach performs better than DARTS, improving the accuracy by 0.28%, 1.64%, 0.34%, 4.5%, and 3.27% compared to DARTS, on CIFAR-10, CIFAR-100, Fashion-MNIST, CompCars, and MIO-TCD datasets, respectively.

Path Following and Collision Avoidance, be it for unmanned surface vessels or other autonomous vehicles, are two fundamental guidance problems in robotics. For many decades, they have been subject to academic study, leading to a vast number of proposed approaches. However, they have mostly been treated as separate problems, and have typically relied on non-linear first-principles models with parameters that can only be determined experimentally. The rise of Deep Reinforcement Learning (DRL) in recent years suggests an alternative approach: end-to-end learning of the optimal guidance policy from scratch by means of a trial-and-error based approach. In this article, we explore the potential of Proximal Policy Optimization (PPO), a DRL algorithm with demonstrated state-of-the-art performance on Continuous Control tasks, when applied to the dual-objective problem of controlling an underactuated Autonomous Surface Vehicle in a COLREGs compliant manner such that it follows an a priori known desired path while avoiding collisions with other vessels along the way. Based on high-fidelity elevation and AIS tracking data from the Trondheim Fjord, an inlet of the Norwegian sea, we evaluate the trained agent's performance in challenging, dynamic real-world scenarios where the ultimate success of the agent rests upon its ability to navigate non-uniform marine terrain while handling challenging, but realistic vessel encounters.

In this paper, we initiate the study of sample complexity of teaching, termed as "teaching dimension" (TDim) in the literature, for Q-learning. While the teaching dimension of supervised learning has been studied extensively, these results do not extend to reinforcement learning due to the temporal constraints posed by the underlying Markov Decision Process environment. We characterize the TDim of Q-learning under different teachers with varying control over the environment, and present matching optimal teaching algorithms. Our TDim results provide the minimum number of samples needed for reinforcement learning, thus complementing standard PAC-style RL sample complexity analysis. Our teaching algorithms have the potential to speed up RL agent learning in applications where a helpful teacher is available.

Noun phrases and relational phrases in Open Knowledge Bases are often not canonical, leading to redundant and ambiguous facts. In this work, we integrate structural information (from which tuple, which sentence) and semantic information (semantic similarity) to do the canonicalization. We represent the two types of information as a multi-layered graph: the structural information forms the links across the sentence, relational phrase, and noun phrase layers; the semantic information forms weighted intra-layer links for each layer. We propose a graph neural network model to aggregate the representations of noun phrases and relational phrases through the multi-layered meta-graph structure. Experiments show that our model outperforms existing approaches on a public datasets in general domain.

Explainable AI (XAI) aims to provide interpretations for predictions made by learning machines, such as deep neural networks, in order to make the machines more transparent for the user and furthermore trustworthy also for applications in e.g. safety-critical areas. So far, however, no methods for quantifying uncertainties of explanations have been conceived, which is problematic in domains where a high confidence in explanations is a prerequisite. We therefore contribute by proposing a new framework that allows to convert any arbitrary explanation method for neural networks into an explanation method for Bayesian neural networks, with an in-built modeling of uncertainties. Within the Bayesian framework a network's weights follow a distribution that extends standard single explanation scores and heatmaps to distributions thereof, in this manner translating the intrinsic network model uncertainties into a quantification of explanation uncertainties. This allows us for the first time to carve out uncertainties associated with a model explanation and subsequently gauge the appropriate level of explanation confidence for a user (using percentiles). We demonstrate the effectiveness and usefulness of our approach extensively in various experiments, both qualitatively and quantitatively.

This competition paper presents microPhantom, a bot playing microRTS and participating in the 2020 microRTS AI competition. microPhantom is based on our previous bot POAdaptive which won the partially observable track of the 2018 and 2019 microRTS AI competitions. In this paper, we focus on decision-making under uncertainty, by tackling the Unit Production Problem with a method based on a combination of Constraint Programming and decision theory. We show that using our method to decide which units to train improves significantly the win rate against the second-best microRTS bot from the partially observable track. We also show that our method is resilient in chaotic environments, with a very small loss of efficiency only. To allow replicability and to facilitate further research, the source code of microPhantom is available, as well as the Constraint Programming toolkit it uses.

Non-stationarity arises in Reinforcement Learning (RL) even in stationary environments. Most RL algorithms collect new data throughout training, using a non-stationary behaviour policy. Furthermore, training targets in RL can change even with a fixed state distribution when the policy, critic, or bootstrap values are updated. We study these types of non-stationarity in supervised learning settings as well as in RL, finding that they can lead to worse generalisation performance when using deep neural network function approximators. Consequently, to improve generalisation of deep RL agents, we propose Iterated Relearning (ITER). ITER augments standard RL training by repeated knowledge transfer of the current policy into a freshly initialised network, which thereby experiences less non-stationarity during training. Experimentally, we show that ITER improves performance on the challenging generalisation benchmarks ProcGen and Multiroom.