AAAI Conferences

Justine Cassell, Professor of Communication Studies and Computer Science, Northwestern University An embodied conversational agent (ECA) is a multimodal interface that adopts some of the properties of human face-to-face conversation, including producing and responding to verbal and nonverbal input and using conversational functions such as turn taking and feedback. ECAs rely on the visual dimension of interacting with an animated character on a screen. Generating conversational behavior for ECAs therefore depends on insights from graphics, vision, speech recognition and synthesis, artificial intelligence, human-computer interface design, and computational linguistics. My work in this interdisciplinary domain (a) highlights the relationship between discourse phenomena and nonverbal behaviors in conversations with computers and between humans mediated by computers, (b) demonstrates their rule-based generativity at a number of levels, and (c) evaluates their effect on human-computer interaction. In this address I will talk about generation of language and graphics (nonverbal behaviors for animated agents) from both underlying concepts and typed text.

What is Machine Behavior?


Understanding the behavior of artificial intelligence(AI) agents is one of the pivotal challenges of the next decade of AI. Interpretability or explainability are some of the terms often used to describe methods that provide insights about the behavior of AI programs. Until today, most of the interpretability techniques have focused on exploring the internal structure of deep neural networks. Recently, a group of AI researchers from the Massachusetts Institute of Technology(MIT) are exploring a radical approach that attempts to explain the behavior of AI observing them in the same we study human or animal behavior. They group the ideas in this area under the catchy name of machine behavior which promises to be one of the most exciting fields in the next few years of AI.

Oh deer: Monkey caught in flagrante delict-doe

The Japan Times

PARIS – Scientists on Tuesday revealed the "highly unusual" behavior of a male monkey filmed trying to have sex with female deer in Japan -- a rare case of inter-species nookie. Sex between animals from different species is uncommon, but exceptional cases are known to occur, chiefly in domesticated and captive animals, scientists reported in the journal Primates. Mating is usually driven by the need to procreate, while sex across the species line is mostly fruitless or yields sterile offspring. For the new study -- only the second on the phenomenon of inter-species sex -- a Japanese macaque, or "snow monkey," was filmed mounting at least two female Sika deer much larger than itself. Without penetration, the young monkey makes sexual movements while riding on the does' backs on Yakushima Island.


AAAI Conferences

Inverse reinforcement learning (IRL) aims to recover the reward function underlying a Markov Decision Process from behaviors of experts in support of decision-making. Most recent work on IRL assumes the same level of trustworthiness of all expert behaviors, and frames IRL as a process of seeking reward function that makes those behaviors appear (near)-optimal. However, it is common in reality that noisy expert behaviors disobeying the optimal policy exist, which may degrade the IRL performance significantly. To address this issue, in this paper, we develop a robust IRL framework that can accurately estimate the reward function in the presence of behavior noise. In particular, we focus on a special type of behavior noise referred to as sparse noise due to its wide popularity in real-world behavior data. To model such noise, we introduce a novel latent variable characterizing the reliability of each expert action and use Laplace distribution as its prior. We then devise an EM algorithm with a novel variational inference procedure in the E-step, which can automatically identify and remove behavior noise in reward learning. Experiments on both synthetic data and real vehicle routing data with noticeable behavior noise show significant improvement of our method over previous approaches in learning accuracy, and also show its power in de-noising behavior data.