Celia Escamilla-Rivera (May 1st 2020). Bayesian Deep Learning for Dark Energy [Online First], IntechOpen, DOI: 10.5772/intechopen.91466. 

Jing Chen and Junying Jia (March 29th 2020). Automatic Recognition of Tea Diseases Based on Deep Learning [Online First], IntechOpen, DOI: 10.5772/intechopen.91953.

Andreas Koch (November 5th 2018). Generating Reality with Geosimulation Models: An Agent-Based Social-Spatial Network Modelling Perspective, Spatial Analysis, Modelling and Planning, Jorge Rocha and José António Tenedório, IntechOpen, DOI: 10.5772/intechopen.77322. 

New Zealand

One of the fundamental challenges of governance is deciding when and how to intervene in multi-agent systems in order to impact group-wide metrics of success. This is particularly challenging when proposed interventions are novel and expensive. For example, one may wish to modify a building’s layout to improve the efficiency of its escape route. Evaluating such interventions would generally require access to an elaborate simulator, which must be constructed ad-hoc for each environment, and can be prohibitively costly or inaccurate. Here we examine a simple alternative: Optimize By Observational Extrapolation (OBOE). The idea is to use observed behavioural trajectories, without any interventions, to learn predictive models mapping environment states to individual agent outcomes, and then use these to evaluate and select changes. We evaluate OBOE in socially complex gridworld environments and consider novel physical interventions that our models were not trained on. We show that neural network models trained to predict agent returns on baseline environments are effective at selecting among the interventions. Thus, OBOE can provide guidance for challenging questions like: “which wall should I tear down in order to minimize the Gini index of this group?”

AAMAS 2020

Should I tear down this wall? Optimizing social metrics by evaluating novel actions

AAMAS is the leading scientific conference for research in autonomous agents and multi-agent systems. The AAMAS conference series was initiated in 2002 as the merging of three respected scientific meetings: the International Conference on Multi-Agent Systems (ICMAS), the International Workshop on Agent Theories, Architectures, and Languages (ATAL), and the International Conference on Autonomous Agents (AA). The aim of the joint conference is to provide a single, high-profile, internationally-respected archival forum for scientific research in the theory and practice of autonomous agents and multi-agent systems.

Browse keynotes, discussions, panels and over 300 presentations.


AAMAS is the leading scientific conference for research in autonomous agents and multi-agent systems. The AAMAS conference series was initiated in 2002 as the merging of three respected scientific meetings: the International Conference on Multi-Agent Systems (ICMAS), the International Workshop on Agent Theories, Architectures, and Languages (ATAL), and the International Conference on Autonomous Agents (AA). 

technical paper

Automatic Classification of Freely-Offered Opinions

Explain how to extract discussion structure which is not abstract

Extracting Discussion Structures for Automated Facilitation Agents

Researchers have long been interested in the role that norms can play in governing agent actions in multi-agent systems. Much work has been done on formalising normative concepts from human society and adapting them for the government of open software systems, and on the simulation of normative processes in human and artificial societies. However, there has been comparatively little work on applying normative MAS mechanisms to understanding the norms in human society. This work investigates this issue in the context of international politics. Using the GDELT dataset, containing machine-encoded records of international events extracted from news reports, we extracted bilateral sequences of inter-country events and applied a Bayesian norm mining mechanism to identify norms that best explained the observed behaviour. A statistical evaluation showed that the normative model fitted the data significantly better than a probabilistic discrete event model.

Mining International Political Norms from the GDELT Database

This paper proposes an approach that aims to extract online discussions structure and to reduce the extraction model size simultaneously. The ultimate goal is the development of an automated facilitation agent that automatically facilitates online discussions. In large-scale online discussions, it becomes necessary to have a facilitator. However, facilitates large-scale online discussions with hundreds or thousands of participants is an intricate problem, and in dynamic discussions, it be- comes impossible for human facilitators to keep an eye on the flow of discussions. As a result, we propose the development of an automated facilitator agent that automatically facilitates online discussions. In order for the automated facilitation agent to function properly, it needs to dynamically extract the structure of online discussions. Towards this end, we use Bidirectional Encoder Representative from Transformers(BERT) and deep neural networks for extracting the structure of online discus- sions. In order to evaluate the proposed approach, we use a dataset based on extracting different posts from a large-scale online discussion system called D-Agree. The experimental results show that the proposed approach is efficient for extracting the structure of online discussions where the link extraction model size is much smaller than the existing methods.

Deep Bidirectional Transformers for Online Discussion Understanding

The enactment of dynamic workflows may take advantage of the multi-agent system paradigm. The approach presented in this pa- per allows exploiting a high-level BPMN definition to generate several agent organisations that can enact the BPMN workflow. The availabil- ity of different solutions (i.e. organisations) enables the optimisation and adaptation features of the approach. The mapping of the initial workflow to organisations starts with the automatic generation of goals from the BPMN, and it exploits a metamodeling approach to generate MOISE organisation definitions.

Goal-driven Adaptation of MOISE Organizations for Workflow Enactment

Many to Many Matching of Agents with Nonlinear Utility Function

Social influence is an essential aspect of human social interaction. Influence propagation modeling has been extensively studied and applied in various research fields, such as the maximization of product adoption, the spread of infectious diseases, etc. The primary sources of information in any social network are the influential nodes that propagate this information further. Our study gives a method to identify the strong influencers in a network and dynamically predict the links/interactions made by them during information propagation. Previously, many efforts have been made to solve the link prediction problem in social networks such as the Adamic Adar index, based on the degree of common connecting neighbors; however, they disregard the fact that social networks can be dynamic over time—any network dataset collected by a snapshot at a particular instant of time may be incomplete and not reflect all the previous links formed. This may in turn result in false predictions. Our approach considers the current dynamics of the social network where information propagated by an agent leads to the prediction of new links with other nodes. Taking opinion propagation as an essential feature, we test our algorithm on Stanford’s Facebook network dataset and compare the results with the previous Adamic Adar index. The results obtained cover all the links predicted by Adamic Adar, as well as some new links formed dynamically. This is indicative that considering dynamic link formation is more efficient and helps address the problem of the incompleteness of network datasets reflecting the state at a time instant.

Dynamic Interactions by Strong Influencers in Social Networks using Opinion Propagation

We present a multi-agent learning algorithm, ALMA-Learning, for efficient and fair allocations in large-scale systems. The proposed approach circumvents the traditional pitfalls of multi-agent learning (e.g., the moving target problem, the curse of dimensionality, or the need for mutually consistent actions) by relying on the ALMA heuristic as a coordination mechanism for each stage game. ALMA-Learning is decentralized and requires no communication between the agents. The latter only observe their own history of action/reward pairs.

Learning to Persist or Switch: Efficient and Fair Allocations in Large-scale Multi-agent Systems

We consider the problem of forming a testbed of elections to be used
for numerical experiments (such as testing algorithms or estimating
the frequency of a given phenomenon). We seek elections that
come from well-known statistical distributions and are as diverse
as possible. To this end, we define a (pseudo)metric over elections,
generate a set of election instances, and measure distances between
them, to assess how diverse they are. Finally, we show how to use
these elections to test election-related algorithms.

Drawing a Map of Elections

"Autonomous agents promise users of a personalized future, allowing
them to direct their attention to tasks most meaningful to them.
However, the demands of personalization stand unfulfilled by current
agent training paradigms such as machine learning, which
require many orders of data to train agents on a single task. In
sequential decision making domains, Reinforcement Learning (RL)
enables this need, when a priori training of desired behaviors is
intractable. Prior work has leveraged user input to train agents
by mapping them to numerical reward signals. However, recent
approaches have identified inconsistent human feedback as a bottleneck
to achieving best-case performance. In this work, we present
empirical evidence to show that human perception affected by
contrast effects distorts their feedback to Reinforcement Learning
agents. Through a set of studies involving 900 participants from
Amazon Mechanical Turk who were asked to give feedback to RL
agents, we show that participants significantly underrate an agent’s
actions after being exposed to an agent of higher competence on the
same task. To understand the significance of this effect on agent performance
during training, we then simulate trainers that underrate
actions of an agent based on past performance – creating a systematically
skewed feedback signal – integrated into an actor-critic
framework. Our results show that agent performance is reduced by
up to 98% in the presence of systematic skews in human feedback in
Atari environments. Our work provides a conceptual understanding
of a source of inconsistency in human feedback, thus informing the
design of human-agent interactions."

Should I tear down this wall? Optimizing social metrics by evaluating novel actions

Downloads

Next from AAMAS 2020

Automatic Classification of Freely-Offered Opinions

Similar lecture

Task Coordination in Multiagent Systems

Should I tear down this wall? Optimizing social metrics by evaluating novel actions

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from AAMAS 2020

Automatic Classification of Freely-Offered Opinions

Similar lecture

Task Coordination in Multiagent Systems

Downloads