A recommender system based on the collaborative behavior of bird flocks

Saka, Esin; Nasraoui, Olfa

doi:10.4108/icst.collaboratecom.2010.11

Cited by 3 publications

(4 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Within the context of online machine learning from human activity data, dynamic usage patterns were studied in [60] and dynamic recommender systems were studied within a stream data mining framework in [59]. In [77], a Swarm Intelligence-based recommender system, inspired by the collaborative behavior of bird flocks and called FlockRecom, generated recommendations by iteratively adjusting the position and speed of dynamic flocks of agents in a virtual space. Even with taking previous work, including all the aforementioned work into account, no prior work has studied human-algorithm interaction via iterated learning mechanisms.…”

Section: Comparison With Existing Work and Limitationsmentioning

confidence: 99%

Human-Algorithm Interaction Biases in the Big Data Cycle: A Markov Chain Iterated Learning Framework

Nasraoui¹,

Shafto²

2016

Preprint

Self Cite

View full text Add to dashboard Cite

Early supervised machine learning algorithms have relied on reliable expert labels to build predictive models. However, the gates of data generation have recently been opened to a wider base of users who started participating increasingly with casual labeling, rating, annotating, etc. The increased online presence and participation of humans has led not only to a democratization of unchecked inputs to algorithms, but also to a wide democratization of the "consumption" of machine learning algorithms' outputs by general users. Hence, these algorithms, many of which are becoming essential building blocks of recommender systems and other information filters, started interacting with users at unprecedented rates. The result is machine learning algorithms that consume more and more data that is unchecked, or at the very least, not fitting conventional assumptions made by various machine learning algorithms. These include biased samples, biased labels, diverging training and testing sets, and cyclical interaction between algorithms, humans, information consumed by humans, and data consumed by algorithms. Yet, the continuous interaction between humans and algorithms is rarely taken into account in machine learning algorithm design and analysis. In this paper, we present a preliminary theoretical model and analysis of the mutual interaction between humans and algorithms, based on an iterated learning framework that is inspired from the study of human language evolution. We also define the concepts of human and algorithm blind spots and outline machine learning approaches to mend iterated bias through two novel notions: antidotes and reactive learning.

show abstract

Section: Comparison With Existing Work and Limitationsmentioning

confidence: 99%

Human-Algorithm Interaction Biases in the Big Data Cycle: A Markov Chain Iterated Learning Framework

Nasraoui¹,

Shafto²

2016

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Some examples of population based algorithm in the related literature are:  Ant colonies [79],  Bird flocks [80],  Fish schools [81].…”

Section: A Particle Swarm Optimisation Methods For Learning Of Fuzzy Cmentioning

confidence: 99%

Tackling Travel Behaviour: An approach based on Fuzzy Cognitive Maps

León¹,

Nápoles²,

Bello³

et al. 2013

IJCIS

View full text Add to dashboard Cite

Although the individuals" transport behavioural modelling is a complex task, it can produce a notable social and economic impact. In this paper, Fuzzy Cognitive Maps are explored to represent the behaviour and operation of such complex systems. An automatic approach to extract mental representations from individuals and convert them into computational structures is defined. For the creation of knowledge bases the use of Knowledge Engineering is accounted and later on the data is transferred into structures based on Fuzzy Cognitive Maps. Once the maps are created, their performances get improved through the use of a Particle Swarm Optimisation algorithm as a learning method, readjusting its predicting capacity from stored scenarios, where individuals left their preferences in front of random situations. Another important result is clustering the maps for knowledge discovery. This permits to find useful groups of individuals that policymakers can use for simulating new rules and policies. After related maps are identified, to merge them as a unique structure could benefit for different usages. Therefore an aggregating procedure is elaborated for this task, constituting an alternative approach for selecting a centroid of a specific estimated group, and therefore having, in only one structure, the knowledge and behavioural acting from a collection of individuals. Learning, clustering and aggregation of Fuzzy Cognitive Maps are combined in a cascade experiment, with the intention of describing travellers" behaviour and change trends in different abstraction levels. The results of this approach will help transportation policy decision makers to understand the people"s needs in a better way, consequently will help them actualising different policy formulations and implementations.

show abstract

“…In such an environment, without continuous or dynamic exploration, the same recommendations would be repeatedly suggested because the system would eventually converge and stagnate on one set of choices. Our initial experimental results show that FlockRecom is a promising approach for recommendation in dynamic environments, thus having potential applications in social networking platforms [130].…”

Section: Summary Of Developed Methodsmentioning

confidence: 95%

“…The results were compared to the traditional user-based nearest neighbor collaborative filtering and FlockRecom was more successful at providing variety in the recommendations without losing recommendation quality [130].…”

Section: Summary Of Contributionsmentioning

confidence: 99%

Swarm intelligence for clustering dynamic data sets for web usage mining and personalization.

Saka¹

Self Cite

View full text Add to dashboard Cite

A recommender system based on the collaborative behavior of bird flocks

Cited by 3 publications

References 14 publications

Human-Algorithm Interaction Biases in the Big Data Cycle: A Markov Chain Iterated Learning Framework

Human-Algorithm Interaction Biases in the Big Data Cycle: A Markov Chain Iterated Learning Framework

Tackling Travel Behaviour: An approach based on Fuzzy Cognitive Maps

Swarm intelligence for clustering dynamic data sets for web usage mining and personalization.

Contact Info

Product

Resources

About