There Is No Agency Without Attention

This paper considers inference over distributed linear Gaussian models using factor graphs and Gaussian belief propagation (BP). The distributed inference algorithm involves only local computation of the information matrix and of the mean vector, and message passing between neighbors. Under broad conditions, it is shown that the message information matrix converges to a unique positive definite limit matrix for arbitrary positive semidefinite initialization, and it approaches an arbitrarily small neighborhood of this limit matrix at a doubly exponential rate. A necessary and sufficient convergence condition for the belief mean vector to converge to the optimal centralized estimator is provided under the assumption that the message information matrix is initialized as a positive semidefinite matrix. Further, it is shown that Gaussian BP always converges when the underlying factor graph is given by the union of a forest and a single loop. The proposed convergence condition in the setup of distributed linear Gaussian models is shown to be strictly weaker than other existing convergence conditions and requirements, including the Gaussian Markov random field based walk-summability condition, and applicable to a large class of scenarios.

This is an area where statistics-and MLbased that cognitive system developers currently address systems can be symbiotic with cognitive systems: the and methodological preferences that they, by and former can provide advanced computation frameworks large, share. For some of the issues, the consensus is while the latter can provide content-related not entirely universal, which is to be expected for a insights into the choice of the inventory of features to group of active developers. Still, the general points of consensus should help to characterize the overall be used in making decisions.

It is also important to remember that having a very sharp distinction of AI A rise in real-world applications of AI has stimulated for social good research is not always feasible, and significant interest from the public, media, and policy often unnecessary. While there has been significant makers. Along with this increasing attention has progress, there still exist many major challenges facing come a media-fueled concern about purported negative the design of effective AIbased approaches to deal consequences of AI, which often overlooks the with the difficulties in real-world domains. One of the societal benefits that AI is delivering and can deliver challenges is interpretability since most algorithms for in the near future. To address these concerns, the AI for social good problems need to be used by human symposium on Artificial Intelligence for the Social end users. Second, the lack of access to valuable data Good (AISOC-17) highlighted the benefits that AI can that could be crucial to the development of appropriate bring to society right now. It brought together AI algorithms is yet another challenge. Third, the researchers and researchers, practitioners, experts, data that we get from the real world is often noisy and and policy makers from a wide variety of domains.

For decades AI researchers have built agents that are capable of carrying out tasks that require human-level or human-like intelligence. During this time, questions of how these programs compared in kind to humans have surfaced and led to beneficial interdisciplinary discussions, but conceptual progress has been slower than technological progress. Within the past decade, the term agency has taken on new import as intelligent agents have become a noticeable part of our everyday lives. Research on autonomous vehicles and personal assistants has expanded into private industry with new and increasingly capable products surfacing as a matter of routine. This wider use of AI technologies has raised questions about legal and moral agency at the highest levels of government (National Science and Technology Council 2016) and drawn the interest of other academic disciplines and the general public. Within this context, the notion of an intelligent agent in AI is too coarse and in need of refinement. We suggest that the space of AI agents can be subdivided into classes, where each class is defined by an associated degree of control.

Developing cognitive agents with human-level natural language understanding (NLU) capabilities requires modeling human cognition because natural, unedited utterances regularly contain ambiguities, ellipses, production errors, implicatures, and many other types of complexities. Moreover, cognitive agents must be nimble in the face of incomplete interpretations since even people do not perfectly understand every aspect of every utterance they hear. So, once an agent has reached the best interpretation it can, it must determine how to proceed – be that acting upon the new information directly, remembering an incomplete interpretation and waiting to see what happens next, seeking out information to fill in the blanks, or asking its interlocutor for clarification. The reasoning needed to support NLU extends far beyond language itself, including, non-exhaustively, the agent’s understanding of its own plans and goals; its dynamic modeling of its interlocutor’s knowledge, plans, and goals, all guided by a theory of mind; its recognition of diverse aspects human behavior, such as affect, cooperative behavior, and the effects of cognitive biases; and its integration of linguistic interpretations with its interpretations of other perceptive inputs, such as simulated vision and non-linguistic audition. Considering all of these needs, it seems hardly possible that fundamental NLU will ever be achieved through the kinds of knowledge-lean text-string manipulation being pursued by the mainstream natural language processing (NLP) community. Instead, it requires a holistic approach to cognitive modeling of the type we are pursuing in a paradigm called OntoAgent.

Morality is a fundamentally human trait which permeates all levels of human society, from basic etiquette and normative expectations of social groups, to formalized legal principles upheld by societies. Hence, future interactive AI systems, in particular, cognitive systems on robots deployed in human settings, will have to meet human normative expectations, for otherwise these system risk causing harm. While the interest in “machine ethics” has increased rapidly in recent years, there are only very few current efforts in the cognitive systems community to investigate moral and ethical reasoning. And there is currently no cognitive architecture that has even rudimentary moral or ethical competence, i.e., the ability to judge situations based on moral principles such as norms and values and make morally and ethically sound decisions. We hence argue for the urgent need to instill moral and ethical competence in all cognitive system intended to be employed in human social contexts.

Neural networks have been successfully implemented in a plethora of prediction tasks ranging from speech interpretation to facial recognition. Because of groundbreaking work in optimization techniques (such as batch normalization, Ioffe and Szegedy (2015)) and model architecture (convolutional, deep belief, and LSTM networks), it is now tractable to use deep neural networks to effectively learn a better feature representation compared to handcrafted methods. 1 One area where such methods have not been utilized is the space of adversarial multiagent systems (for example, multiple independent players in competition), specifically when the multiagent behavior comes in the form of trajectories. There are two reasons for this: i) procuring large volumes of data where deep methods are effective is difficult to obtain, and ii) forming an initial representation of the raw trajectories so that deep neural networks are effective is challenging. In this paper, we explore the effectiveness of deep neural networks on a large volume of basketball tracking data, which contains the x, y locations of multiple agents (players) in an adversarial domain (game). To thoroughly explore this problem, we focus on the following task: "given the trajectories of the players and ball in the previous five seconds, can we accurately predict the likelihood that a player with position/role X will make the shot?"

Recent progress in artificial intelligence, especially in the area of deep learning, has been breath-taking. This is very encouraging for anyone interested in the field, yet the true progress towards human-level artificial intelligence is much harder to evaluate. The evaluation of artificial intelligence is a very difficult problem for a number of reasons. For example, the lack of consensus on the basic desiderata necessary for intelligent machines is one of the primary barriers to the development of unified approaches towards comparing different agents. José Hernández-Orallo or Kristinn R. Thórisson to name a few), the area would benefit from more attention from the AI community.

A new amortized variance-reduced gradient (AVRG) algorithm was developed in [1], which has constant storage requirement in comparison to SAGA and balanced gradient computations in comparison to SVRG. One key advantage of the AVRG strategy is its amenability to decentralized implementations. In this work, we show how AVRG can be extended to the network case where multiple learning agents are assumed to be connected by a graph topology. In this scenario, each agent observes data that is spatially distributed and all agents are only allowed to communicate with direct neighbors. Moreover, the amount of data observed by the individual agents may differ drastically. For such situations, the balanced gradient computation property of AVRG becomes a real advantage in reducing idle time caused by unbalanced local data storage requirements, which is characteristic of other reduced-variance gradient algorithms. The resulting diffusion-AVRG algorithm is shown to have linear convergence to the exact solution, and is much more memory efficient than other alternative algorithms. In addition, by using a mini-batch strategy, it is shown that diffusion-AVRG is more computationally efficient than exact diffusion or EXTRA while maintaining almost the same amount of communications.