P. M. Krafft

P. M. Krafft, Meg Young, Michael Katell et al.

Recent concern about harms of information technologies motivate consideration of regulatory action to forestall or constrain certain developments in the field of artificial intelligence (AI). However, definitional ambiguity hampers the possibility of conversation about this urgent topic of public concern. Legal and regulatory interventions require agreed-upon definitions, but consensus around a definition of AI has been elusive, especially in policy conversations. With an eye towards practical working definitions and a broader understanding of positions on these issues, we survey experts and review published policy documents to examine researcher and policy-maker conceptions of AI. We find that while AI researchers favor definitions of AI that emphasize technical functionality, policy-makers instead use definitions that compare systems to human thinking and behavior. We point out that definitions adhering closely to the functionality of AI systems are more inclusive of technologies in use today, whereas definitions that emphasize human-like capabilities are most applicable to hypothetical future technologies. As a result of this gap, ethical and regulatory efforts may overemphasize concern about future technologies at the expense of pressing issues with existing deployed technologies.

Michael Katell, Meg Young, Bernease Herman et al.

A wave of recent scholarship documenting the discriminatory harms of algorithmic systems has spurred widespread interest in algorithmic accountability and regulation. Yet effective accountability and regulation is stymied by a persistent lack of resources supporting public understanding of algorithms and artificial intelligence. Through interactions with a US-based civil rights organization and their coalition of community organizations, we identify a need for (i) heuristics that aid stakeholders in distinguishing between types of analytic and information systems in lay language, and (ii) risk assessment tools for such systems that begin by making algorithms more legible. The present work delivers a toolkit to achieve these aims. This paper both presents the Algorithmic Equity Toolkit (AEKit) Equity as an artifact, and details how our participatory process shaped its design. Our work fits within human-computer interaction scholarship as a demonstration of the value of HCI methods and approaches to problems in the area of algorithmic transparency and accountability.

Yea-Seul Kim, Logan A Walls, Peter Krafft et al.

People naturally bring their prior beliefs to bear on how they interpret the new information, yet few formal models exist for accounting for the influence of users' prior beliefs in interactions with data presentations like visualizations. We demonstrate a Bayesian cognitive model for understanding how people interpret visualizations in light of prior beliefs and show how this model provides a guide for improving visualization evaluation. In a first study, we show how applying a Bayesian cognition model to a simple visualization scenario indicates that people's judgments are consistent with a hypothesis that they are doing approximate Bayesian inference. In a second study, we evaluate how sensitive our observations of Bayesian behavior are to different techniques for eliciting people subjective distributions, and to different datasets. We find that people don't behave consistently with Bayesian predictions for large sample size datasets, and this difference cannot be explained by elicitation technique. In a final study, we show how normative Bayesian inference can be used as an evaluation framework for visualizations, including of uncertainty.

Dhaval Adjodah, Dan Calacci, Abhimanyu Dubey et al.

In this empirical paper, we investigate how learning agents can be arranged in more efficient communication topologies for improved learning. This is an important problem because a common technique to improve speed and robustness of learning in deep reinforcement learning and many other machine learning algorithms is to run multiple learning agents in parallel. The standard communication architecture typically involves all agents intermittently communicating with each other (fully connected topology) or with a centralized server (star topology). Unfortunately, optimizing the topology of communication over the space of all possible graphs is a hard problem, so we borrow results from the networked optimization and collective intelligence literatures which suggest that certain families of network topologies can lead to strong improvements over fully-connected networks. We start by introducing alternative network topologies to DRL benchmark tasks under the Evolution Strategies paradigm which we call Network Evolution Strategies. We explore the relative performance of the four main graph families and observe that one such family (Erdos-Renyi random graphs) empirically outperforms all other families, including the de facto fully-connected communication topologies. Additionally, the use of alternative network topologies has a multiplicative performance effect: we observe that when 1000 learning agents are arranged in a carefully designed communication topology, they can compete with 3000 agents arranged in the de facto fully-connected topology. Overall, our work suggests that distributed machine learning algorithms would learn more efficiently if the communication topology between learning agents was optimized.

Daniel Y. Fu, Emily S. Wang, Peter M. Krafft et al.

Flocking is a coordinated collective behavior that results from local sensing between individual agents that have a tendency to orient towards each other. Flocking is common among animal groups and might also be useful in robotic swarms. In the interest of learning how to control flocking behavior, recent work in the multiagent systems literature has explored the use of influencing agents for guiding flocking agents to face a target direction. The existing work in this domain has focused on simulation settings of small areas with toroidal shapes. In such settings, agent density is high, so interactions are common, and flock formation occurs easily. In our work, we study new environments with lower agent density, wherein interactions are more rare. We study the efficacy of placement strategies and influencing agent behaviors drawn from the literature, and find that the behaviors that have been shown to work well in high-density conditions tend to be much less effective in lower density environments. The source of this ineffectiveness is that the influencing agents explored in prior work tended to face directions optimized for maximal influence, but which actually separate the influencing agents from the flock. We find that in low-density conditions maintaining a connection to the flock is more important than rushing to orient towards the desired direction. We use these insights to propose new influencing agent behaviors, which we dub "follow-then-influence"; agents act like normal members of the flock to achieve positions that allow for control and then exert their influence. This strategy overcomes the difficulties posed by low density environments.

Peter M Krafft, Nicolás Della Penna, Alex Pentland

As cryptocurrencies gain popularity and credibility, marketplaces for cryptocurrencies are growing in importance. Understanding the dynamics of these markets can help to assess how viable the cryptocurrnency ecosystem is and how design choices affect market behavior. One existential threat to cryptocurrencies is dramatic fluctuations in traders' willingness to buy or sell. Using a novel experimental methodology, we conducted an online experiment to study how susceptible traders in these markets are to peer influence from trading behavior. We created bots that executed over one hundred thousand trades costing less than a penny each in 217 cryptocurrencies over the course of six months. We find that individual "buy" actions led to short-term increases in subsequent buy-side activity hundreds of times the size of our interventions. From a design perspective, we note that the design choices of the exchange we study may have promoted this and other peer influence effects, which highlights the potential social and economic impact of HCI in the design of digital institutions.

Dhaval Adjodah, Dan Calacci, Yan Leng et al.

We draw upon a previously largely untapped literature on human collective intelligence as a source of inspiration for improving deep learning. Implicit in many algorithms that attempt to solve Deep Reinforcement Learning (DRL) tasks is the network of processors along which parameter values are shared. So far, existing approaches have implicitly utilized fully-connected networks, in which all processors are connected. However, the scientific literature on human collective intelligence suggests that complete networks may not always be the most effective information network structures for distributed search through complex spaces. Here we show that alternative topologies can improve deep neural network training: we find that sparser networks learn higher rewards faster, leading to learning improvements at lower communication costs.

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

We study a distributed learning process observed in human groups and other social animals. This learning process appears in settings in which each individual in a group is trying to decide over time, in a distributed manner, which option to select among a shared set of options. Specifically, we consider a stochastic dynamics in a group in which every individual selects an option in the following two-step process: (1) select a random individual and observe the option that individual chose in the previous time step, and (2) adopt that option if its stochastic quality was good at that time step. Various instantiations of such distributed learning appear in nature, and have also been studied in the social science literature. From the perspective of an individual, an attractive feature of this learning process is that it is a simple heuristic that requires extremely limited computational capacities. But what does it mean for the group -- could such a simple, distributed and essentially memoryless process lead the group as a whole to perform optimally? We show that the answer to this question is yes -- this distributed learning is highly effective at identifying the best option and is close to optimal for the group overall. Our analysis also gives quantitative bounds that show fast convergence of these stochastic dynamics. Prior to our work the only theoretical work related to such learning dynamics has been either in deterministic special cases or in the asymptotic setting. Finally, we observe that our infinite population dynamics is a stochastic variant of the classic multiplicative weights update (MWU) method. Consequently, we arrive at the following interesting converse: the learning dynamics on a finite population considered here can be viewed as a novel distributed and low-memory implementation of the classic MWU method.

Peter M Krafft, Michael Macy, Alex Pentland

The use of bots as virtual confederates in online field experiments holds extreme promise as a new methodological tool in computational social science. However, this potential tool comes with inherent ethical challenges. Informed consent can be difficult to obtain in many cases, and the use of confederates necessarily implies the use of deception. In this work we outline a design space for bots as virtual confederates, and we propose a set of guidelines for meeting the status quo for ethical experimentation. We draw upon examples from prior work in the CSCW community and the broader social science literature for illustration. While a handful of prior researchers have used bots in online experimentation, our work is meant to inspire future work in this area and raise awareness of the associated ethical issues.

Peter M. Krafft, Julia Zheng, Wei Pan et al.

Collective intelligence is believed to underly the remarkable success of human society. The formation of accurate shared beliefs is one of the key components of human collective intelligence. How are accurate shared beliefs formed in groups of fallible individuals? Answering this question requires a multiscale analysis. We must understand both the individual decision mechanisms people use, and the properties and dynamics of those mechanisms in the aggregate. As of yet, mathematical tools for such an approach have been lacking. To address this gap, we introduce a new analytical framework: We propose that groups arrive at accurate shared beliefs via distributed Bayesian inference. Distributed inference occurs through information processing at the individual level, and yields rational belief formation at the group level. We instantiate this framework in a new model of human social decision-making, which we validate using a dataset we collected of over 50,000 users of an online social trading platform where investors mimic each others' trades using real money in foreign exchange and other asset markets. We find that in this setting people use a decision mechanism in which popularity is treated as a prior distribution for which decisions are best to make. This mechanism is boundedly rational at the individual level, but we prove that in the aggregate implements a type of approximate "Thompson sampling"---a well-known and highly effective single-agent Bayesian machine learning algorithm for sequential decision-making. The perspective of distributed Bayesian inference therefore reveals how collective rationality emerges from the boundedly rational decision mechanisms people use.

L. Elisa Celis, Peter M. Krafft, Nathan Kobe

One goal of online social recommendation systems is to harness the wisdom of crowds in order to identify high quality content. Yet the sequential voting mechanisms that are commonly used by these systems are at odds with existing theoretical and empirical literature on optimal aggregation. This literature suggests that sequential voting will promote herding---the tendency for individuals to copy the decisions of others around them---and hence lead to suboptimal content recommendation. Is there a problem with our practice, or a problem with our theory? Previous attempts at answering this question have been limited by a lack of objective measurements of content quality. Quality is typically defined endogenously as the popularity of content in absence of social influence. The flaw of this metric is its presupposition that the preferences of the crowd are aligned with underlying quality. Domains in which content quality can be defined exogenously and measured objectively are thus needed in order to better assess the design choices of social recommendation systems. In this work, we look to the domain of education, where content quality can be measured via how well students are able to learn from the material presented to them. Through a behavioral experiment involving a simulated massive open online course (MOOC) run on Amazon Mechanical Turk, we show that sequential voting systems can surface better content than systems that elicit independent votes.

Peter M. Krafft, Chris L. Baker, Alex Pentland et al.

Whether in groups of humans or groups of computer agents, collaboration is most effective between individuals who have the ability to coordinate on a joint strategy for collective action. However, in general a rational actor will only intend to coordinate if that actor believes the other group members have the same intention. This circular dependence makes rational coordination difficult in uncertain environments if communication between actors is unreliable and no prior agreements have been made. An important normative question with regard to coordination in these ad hoc settings is therefore how one can come to believe that other actors will coordinate, and with regard to systems involving humans, an important empirical question is how humans arrive at these expectations. We introduce an exact algorithm for computing the infinitely recursive hierarchy of graded beliefs required for rational coordination in uncertain environments, and we introduce a novel mechanism for multiagent coordination that uses it. Our algorithm is valid in any environment with a finite state space, and extensions to certain countably infinite state spaces are likely possible. We test our mechanism for multiagent coordination as a model for human decisions in a simple coordination game using existing experimental data. We then explore via simulations whether modeling humans in this way may improve human-agent collaboration.