In the world of online gaming, not all actions are created equal. For example, when a player's character is confronted with a closed door, it would not make much sense to brandish a weapon, apply a healing potion, or attempt to barter. A more reasonable response would be to either open or unlock the door. The term interaction mode embodies the idea that many potential actions are neither useful nor applicable in a given situation. This paper presents a AEGIM, an algorithm for the automated extraction of game interaction modes via a semantic embedding space. AEGIM uses an image captioning system in conjunction with a semantic vector space model to create a gestalt representation of in-game screenshots, thus enabling it to detect the interaction mode evoked by the game.

We address the question of how to build AI agents that behave ethically by appealing to a computational creativity framework in which output artifacts are agent behaviors and candidate behaviors are evaluated using a normative ethics as the aesthetic measure. We then appeal again to computational creativity to address the meta-level question of which normative ethics the system should employ as its aesthetic, where now output meta-artifacts are normative ethics and candidate ethics are evaluated using a meta-ethics-based aesthetic. We consider briefly some of the issues raised by such a proposal as well as how the hybrid base-meta-level system might be evaluated from three different perspectives: creative, behavioral and ethical.

Research has shown that a person's financial success is more dependent on the ability to deal with people than on professional knowledge. Sage advice, such as "if you can't say something nice, don't say anything at all" and principles articulated in Carnegie's classic "How to Win Friends and Influence People," offer trusted rules-of-thumb for how people can successfully deal with each other. However, alternative philosophies for dealing with people have also emerged. The success of an AI system is likewise contingent on its ability to win friends and influence people. In this paper, we study how AI systems should be designed to win friends and influence people in repeated games with cheap talk (RGCTs). We create several algorithms for playing RGCTs by combining existing behavioral strategies (what the AI does) with signaling strategies (what the AI says) derived from several competing philosophies. Via user study, we evaluate these algorithms in four RGCTs. Our results suggest sufficient properties for AIs to win friends and influence people in RGCTs.

Visual style transfer involves combining the content of one image with the style of another, and recent work has produced some compelling results. This paper proposes a related task that requires additional system intelligence and autonomy—that of style creation. Rather than using the style of an existing source image, the goal is to have the system autonomously create a rendering style based on a simple (text- based) semantic description. Results demonstrate the system’s ability to autonomously create interesting, semantically appropriate styles that can be applied for image rendering.

We examine the question of whether we can automatically classify the sentiment of individual tweets in Farsi, to determine their changing sentiments over time toward a number of trending political topics. Examining tweets in Farsi adds challenges such as the lack of a sentiment lexicon and part-of-speech taggers, frequent use of colloquial words, and unique orthography and morphology characteristics. We have collected over 1 million Tweets on political topics in the Farsi language, with an annotated data set of over 3,000 tweets. We find that an SVM classifier with Brown clustering for feature selection yields a median accuracy of 56% and accuracy as high as 70%. We use this classifier to track dynamic sentiment during a key period of Irans negotiations over its nuclear program.

When dealing with images and semantics, most computational systems attempt to automatically extract meaning from images. Here we attempt to go the other direction and autonomously create images that communicate concepts. We present an enhanced semantic model that is used to generate novel images that convey meaning. We employ a vector space model and a large corpus to learn vector representations of words and then train the semantic model to predict word vectors that could describe a given image. Once trained, the model autonomously guides the process of rendering images that convey particular concepts. A significant contribution is that, because of the semantic associations encoded in these word vectors, we can also render images that convey concepts on which the model was not explicitly trained. We evaluate the semantic model with an image clustering technique and demonstrate that the model is successful in creating images that communicate semantic relationships.

We present a system for autonomously generating puzzles in the form of a 2D, tile-based world.  Puzzle design is entirely dependent on tile characteristics, which are implemented as abstract classes that can be modified by the system.  Thus, the system controls not only the base-level puzzle design but also (to some extent) the meta-level component design.  The result is a rich space of possible puzzles that the system explores with a combination of evolutionary computation and Q -learning.  The system autonomously produces a variety of puzzles of varying difficulty to create a game called Loki's Castle .  The system is almost completely autonomous, requiring only a minimal description of what a puzzle should include, and the abstraction allows extensibility so that future versions can invent entirely new classes of tiles.  Several puzzle examples are presented to demonstrate the system's capability.

Sum-product networks (SPNs) are rooted, directed acyclic graphs (DAGs) of sum and product nodes with well-defined probabilistic semantics. Moreover, exact inference in the distribution represented by an SPN is guaranteed to take linear time in the size of the DAG. In this paper we introduce an algorithm that learns the structure of an SPN using a greedy search approach. It incorporates methods used in a previous SPN structure-learning algorithm, but, unlike the previous algorithm, is not limited to learning tree-structured SPNs. Several proven ideas from circuit complexity theory along with our experimental results provide evidence for the advantages of SPNs with less-restrictive, non-tree structures.

Many robotic tasks require solutions that maximize multiple performance objectives. For example, in path-planning, these objectives often include finding short paths that avoid risk and maximize the information obtained by the robot. Although there exist many algorithms for multiobjective optimization, few of these algorithms apply directly to robotic path-planning and fewer still are capable of finding the set of Pareto optimal solutions. We present the MORRF*(Multi-Objective Rapidly exploring Random Forest*) algorithm, which blends concepts from two different types of algorithms from the literature: Optimal rapidly exploring random tree (RRT*) for efficient path finding and a decomposition-based approach to multi-objective optimization. The random forest uses two types of tree structures: a set of reference trees and a set of subproblem trees. We present a theoretical analysis that demonstrates that the algorithm asymptotically produces the set of Pareto optimal solutions, and use simulations to demonstrate the effectiveness and efficiency of MORRF* in approximating the Pareto set.

Machine reading is a relatively new field that features computer programs designed to read flowing text and extract fact assertions expressed by the narrative content. This task involves two core technologies: natural language processing (NLP) and information extraction (IE). In this paper we describe a machine reading system that we have developed within a cognitive architecture. We show how we have integrated into the framework several levels of knowledge for a particular domain, ideas from cognitive semantics and construction grammar, plus tools from prior NLP and IE research. The result is a system that is capable of reading and interpreting complex and fairly idiosyncratic texts in the family history domain. We describe the architecture and performance of the system. After presenting the results from several evaluations that we have carried out, we summarize possible future directions.