To solve tasks in new environments involving objects unseen during training, agents must reason over prior information about those objects and their relations. We introduce the Prior Knowledge Graph network, an architecture for combining prior information, structured as a knowledge graph, with a symbolic parsing of the visual scene, and demonstrate that this approach is able to apply learned relations to novel objects whereas the baseline algorithms fail. Ablation experiments show that the agents ground the knowledge graph relations to semantically-relevant behaviors. In both a Sokoban game and the more complex Pacman environment, our network is also more sample efficient than the baselines, reaching the same performance in 5-10x fewer episodes. Once the agents are trained with our approach, we can manipulate agent behavior by modifying the knowledge graph in semantically meaningful ways. These results suggest that our network provides a framework for agents to reason over structured knowledge graphs while still leveraging gradient based learning approaches.

Systems for automatic argument generation and debate require the ability to (1) determine the stance of any claims employed in the argument and (2) assess the specificity of each claim relative to the argument context. Existing work on understanding claim specificity and stance, however, has been limited to the study of argumentative structures that are relatively shallow, most often consisting of a single claim that directly supports or opposes the argument thesis. In this paper, we tackle these tasks in the context of complex arguments on a diverse set of topics. In particular, our dataset consists of manually curated argument trees for 741 controversial topics covering 95,312 unique claims; lines of argument are generally of depth 2 to 6. We find that as the distance between a pair of claims increases along the argument path, determining the relative specificity of a pair of claims becomes easier and determining their relative stance becomes harder.

Existing argumentation datasets have succeeded in allowing researchers to develop computational methods for analyzing the content, structure and linguistic features of argumentative text. They have been much less successful in fostering studies of the effect of "user" traits -- characteristics and beliefs of the participants -- on the debate/argument outcome as this type of user information is generally not available. This paper presents a dataset of 78, 376 debates generated over a 10-year period along with surprisingly comprehensive participant profiles. We also complete an example study using the dataset to analyze the effect of selected user traits on the debate outcome in comparison to the linguistic features typically employed in studies of this kind.

Public debate forums provide a common platform for exchanging opinions on a topic of interest. While recent studies in natural language processing (NLP) have provided empirical evidence that the language of the debaters and their patterns of interaction play a key role in changing the mind of a reader, research in psychology has shown that prior beliefs can affect our interpretation of an argument and could therefore constitute a competing alternative explanation for resistance to changing one's stance. To study the actual effect of language use vs. prior beliefs on persuasion, we provide a new dataset and propose a controlled setting that takes into consideration two reader level factors: political and religious ideology. We find that prior beliefs affected by these reader level factors play a more important role than language use effects and argue that it is important to account for them in NLP studies of persuasion.

The quantum approximate optimization algorithm (QAOA) has rapidly become a cornerstone of contemporary quantum algorithm development. Despite a growing range of applications, only a few results have been developed towards understanding the algorithms ultimate limitations. Here we report that QAOA exhibits a strong dependence on a problem instances constraint to variable ratio - this problem density places a limiting restriction on the algorithms capacity to minimize a corresponding objective function (and hence solve optimization problems). Such 'reachability deficits' persist even in the absence of barren plateaus [McClean et al., 2018] and are outside of the recently reported level-1 QAOA limitations [Hastings 2019]. Building on general numerical experiments, we compare the presence of reachability deficits with analytic solutions of the variational model of Grover's search algorithm.

Histopathology image analysis can be considered as a Multiple instance learning (MIL) problem, where the whole slide histopathology image (WSI) is regarded as a bag of instances (i.e., patches) and the task is to predict a single class label to the WSI. However, in many reallife applications such as computational pathology, discovering the key instances that trigger the bag label is of great interest because it provides reasons for the decision made by the system. In this paper, we propose a deep convolutional neural network (CNN) model that addresses the primary task of a bag classification on a histopathology image and also learns to identify the response of each instance to provide interpretable results to the final prediction. We incorporate the attention mechanism into the proposed model to operate the transformation of instances and learn attention weights to allow us to find key patches. To perform a balanced training, we introduce adaptive weighing in each training bag to explicitly adjust the weight distribution in order to concentrate more on the contribution of hard samples. Based on the learned attention weights, we further develop a solution to boost the classification performance by generating the bags with hard negative instances. We conduct extensive experiments on colon and breast cancer histopathology data and show that our framework achieves state-of-the-art performance.

Ontology-based knowledge bases (KBs) like DBpedia are very valuable resources, but their usefulness and usability is limited by various quality issues. One such issue is the use of string literals instead of semantically typed entities. In this paper we study the automated canonicalization of such literals, i.e., replacing the literal with an existing entity from the KB or with a new entity that is typed using classes from the KB. We propose a framework that combines both reasoning and machine learning in order to predict the relevant entities and types, and we evaluate this framework against state-of-the-art baselines for both semantic typing and entity matching.

Tool-use applications in robotics require conceptual knowledge about objects for informed decision making and object interactions. State-of-the-art methods employ hand-crafted symbolic knowledge which is defined from a human perspective and grounded into sensory data afterwards. However, due to different sensing and acting capabilities of robots, their conceptual understanding of objects must be generated from a robot's perspective entirely, which asks for robot-centric conceptual knowledge about objects. With this goal in mind, this article motivates that such knowledge should be based on physical and functional properties of objects. Consequently, a selection of ten properties is defined and corresponding extraction methods are proposed. This multi-modal property extraction forms the basis on which our second contribution, a robot-centric knowledge generation is build on. It employs unsupervised clustering methods to transform numerical property data into symbols, and Bivariate Joint Frequency Distributions and Sample Proportion to generate conceptual knowledge about objects using the robot-centric symbols. A preliminary implementation of the proposed framework is employed to acquire a dataset comprising physical and functional property data of 110 houshold objects. This Robot-Centric dataSet (RoCS) is used to evaluate the framework regarding the property extraction methods, the semantics of the considered properties within the dataset and its usefulness in real-world applications such as tool substitution.

Multiagent decision-making problems in partially observable environments are usually modeled as either extensive-form games (EFGs) within the game theory community or partially observable stochastic games (POSGs) within the reinforcement learning community. While most practical problems can be modeled in both formalisms, the communities using these models are mostly distinct with little sharing of ideas or advances. The last decade has seen dramatic progress in algorithms for EFGs, mainly driven by the challenge problem of poker. We have seen computational techniques achieving super-human performance, some variants of poker are essentially solved, and there are now sound local search algorithms which were previously thought impossible. While the advances have garnered attention, the fundamental advances are not yet understood outside the EFG community. This can be largely explained by the starkly different formalisms between the game theory and reinforcement learning communities and, further, by the unsuitability of the original EFG formalism to make the ideas simple and clear. This paper aims to address these hindrances, by advocating a new unifying formalism, a variant of POSGs, which we call Factored-Observation Games (FOGs). We prove that any timeable perfect-recall EFG can be efficiently modeled as a FOG as well as relating FOGs to other existing formalisms. Additionally, a FOG explicitly identifies the public and private components of observations, which is fundamental to the recent EFG breakthroughs. We conclude by presenting the two building-blocks of these breakthroughs --- counterfactual regret minimization and public state decomposition --- in the new formalism, illustrating our goal of a simpler path for sharing recent advances between game theory and reinforcement learning community.

This paper aims to question the suitability of the Turing Test, for testing machine intelligence, in the light of advances made in the last 60 years in science, medicine, and philosophy of mind. While the main concept of the test may seem sound and valid, a detailed analysis of what is required to pass the test highlights a significant flow. Once the analysis of the test is presented, a systematic approach is followed in analysing what is needed to devise a test or tests for intelligent machines. The paper presents a plausible generic framework based on categories of factors implied by subjective perception of intelligence. An evaluative discussion concludes the paper highlighting some of the unaddressed issues within this generic framework.