A Distributed Learning Dynamics in Social Groups

Celis, L. Elisa; Krafft, Peter M.; Vishnoi, Nisheeth K.

doi:10.1145/3087801.3087820

Cited by 12 publications

(19 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, we employ arguments based on the linearity of expectation to handle the dependencies of the regret among the agents induced by our algorithm (cf. Propositions 11,13,16), which we believe can be useful in studying other algorithms for our model.…”

Section: Resultsmentioning

confidence: 95%

“…The graph encodes interactions where 'nearby users' on the graph have 'similar' contextual bandit instances, different from interactions in our model. Recent works [51], [16] have considered the social learning problem where agents do bestarm identification (simple regret). In these setups, the memory of an agent is limited, and hence standard bandit algorithms such as UCB is infeasible.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Social Learning in Multi Agent Multi Armed Bandits

SankararamanAbishek

GaneshAyalvadi

Shakkottai

2019

Proc. ACM Meas. Anal. Comput. Syst.

View full text Add to dashboard Cite

Motivated by emerging need of learning algorithms for large scale networked and decentralized systems, we introduce a distributed version of the classical stochastic Multi-Arm Bandit (MAB) problem. Our setting consists of a large number of agents n that collaboratively and simultaneously solve the same instance of K armed MAB to minimize the average cumulative regret over all agents. The agents can communicate and collaborate among each other only through a pairwise asynchronous gossip based protocol that exchange a limited number of bits. In our model, agents at each point decide on (i) which arm to play, (ii) whether to, and if so (iii) what and whom to communicate with. Agents in our model are decentralized, namely their actions only depend on their observed history in the past.We develop a novel algorithm in which agents, whenever they choose, communicate only armids and not samples, with another agent chosen uniformly and independently at random. The per-agent regret scaling achieved by our algorithm is OFurthermore, any agent in our algorithm communicates (arm-ids to an uniformly and independently chosen agent) only a total of Θ(log(T )) times over a time interval of T . We compare our results to two benchmarks -one where there is no communication among agents and one corresponding to complete interaction, where an agent has access to the entire system history of arms played and rewards obtained of all agents. We show both theoretically and empirically, that our algorithm experiences a significant reduction both in per-agent regret when compared to the case when agents do not collaborate and each agent is playing the standard MAB problem (where regret would scale linearly in K), and in communication complexity when compared to the full interaction setting which requires T communication attempts by an agent over T arm pulls.

show abstract

Section: Resultsmentioning

confidence: 95%

Section: Related Workmentioning

confidence: 99%

Social Learning in Multi Agent Multi Armed Bandits

SankararamanAbishek

GaneshAyalvadi

Shakkottai

2019

Proc. ACM Meas. Anal. Comput. Syst.

View full text Add to dashboard Cite

show abstract

“…(Note, however, that the update in Equation ( 7) is not Bayesian). Our update rule also bears resemblance to the multiplicative weights update method [4] that has been successfully used to explain learning in social groups [13] and sexual evolution [17].…”

Section: Discussion Of the Modelmentioning

confidence: 99%

The Effect of the Rooney Rule on Implicit Bias in the Long Term

Celis¹,

Hays²,

Mehrotra³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

A robust body of evidence demonstrates the adverse effects of implicit bias in various contexts, including hiring, admissions, criminal justice, and healthcare. The Rooney Rule is a simple intervention developed to counter implicit bias in hiring, and has been implemented in the private and public sector in various settings [26,19]. The Rooney Rule requires that a selection panel include at least one candidate from an underrepresented group in their shortlist of candidates. Recently, Kleinberg and Raghavan [42] proposed a mathematical model of implicit bias and studied the effectiveness of the Rooney Rule when applied to a single selection decision. However, selection decisions often occur repeatedly over time; e.g., a software firm is continuously hiring employees, or a university makes admissions decisions every year. Further, it has been observed that, given consistent counterstereotypical feedback, implicit biases against underrepresented candidates can change (e.g., [23]).In this paper, building on Kleinberg and Raghavan's model and work on opinion dynamics, we consider a model of how a selection panel's implicit bias changes over time given their hiring decisions either with or without the Rooney Rule in place. Our main result is that, for this model, when the selection panel is constrained by the Rooney Rule, their implicit bias roughly reduces at a rate that is inverse of the size of the shortlist-independent of the total number of candidates, whereas without the Rooney Rule, the rate is inversely proportional to the number of candidates. Thus, our model predicts that when the number of candidates is much larger than the size of the shortlist, the Rooney Rule enables a significantly faster reduction in implicit bias, providing additional reason in favor of instating it as a strategy to mitigate implicit bias. Towards empirically evaluating the long-term effect of the Rooney Rule in repeated selection decisions, we conduct an iterative candidate selection experiment on Amazon Mechanical Turk. We observe that, indeed, decision-makers subject to the Rooney Rule select more minority candidates in addition to those required by the rule itself than they would if no rule is in effect, and in fact are able to do so without considerably decreasing the utility of candidates selected.

show abstract

“…A simple "social sampling" behavior provides a motivating example of one such mechanism [11,26]. The social sampling procedure consists of people who are undecided about their beliefs sampling proposed beliefs uniformly at random from people who are decided, and then assessing whether to accept the received proposed belief according to evidence.…”

Section: Rumors In Proportion To Evidencementioning

confidence: 99%

Keeping Rumors in Proportion

Krafft

Spiro

2019

Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems

Self Cite

View full text Add to dashboard Cite

The study of rumors has garnered wider attention as regulators and researchers turn towards problems of misinformation on social media. One goal has been to discover and implement mechanisms that promote healthy information ecosystems. Classically defined as regarding ambiguous situations, rumors pose the unique difficulty of intrinsic uncertainty around their veracity. Further complicating matters, rumors can serve the public when they do spread valuable true information. To address these challenges, we develop an approach that reifies "rumor proportions" as central to the theory of systems for managing rumors. We use this lens to advocate for systems that, rather than aiming to stifle rumors entirely or aiming to stop only false rumors, aim to prevent rumors from growing out of proportion relative to normative benchmark representations of intrinsic uncertainty.

show abstract

A Distributed Learning Dynamics in Social Groups

Cited by 12 publications

References 48 publications

Social Learning in Multi Agent Multi Armed Bandits

Social Learning in Multi Agent Multi Armed Bandits

The Effect of the Rooney Rule on Implicit Bias in the Long Term

Keeping Rumors in Proportion

Contact Info

Product

Resources

About