Person:
Zou, James

Profile Picture

Email Address

AA Acceptance Date

Birth Date

Research Projects

Organizational Units

Job Title

Last Name

Zou

First Name

James

Name

Zou, James
Author's Publications: search.filters.isAuthorOfPublication.8b862b1d-a8f6-4fab-977f-9a51670468d2

Search Results

Now showing 1 - 9 of 9
  • Thumbnail Image
    Publication
    Approval Voting Behavior in Doodle Polls
    (2014) Zou, James; Meir, Reshef; Parkes, David
    Doodle is a simple and popular online system for scheduling events. It is an implementation of the approval voting mechanism, where candidates are the time slots and each responder approves a subset of the slots. We analyze all the Doodle polls created in the US from JulySeptember 2011 (over 340,000 polls), consisting of both hidden polls (where you cannot see other people’s votes) and open polls (where you can see all the previous responses). By analyzing the differences in behavior in hidden and open polls, we gain unique insights into strategies that people apply in natural voting settings. Responders in open polls are more likely to approve slots that are very popular or very unpopular, but not intermediate slots. We show that this behavior is inconsistent with models that have been proposed in the voting literature, and propose a new model based on combining personal and social utilities to explain the data.
  • Thumbnail Image
    Publication
    Algorithms and Models for Genome Biology
    (2014-02-25) Zou, James; Parkes, David C.; Brenner, Michael; Bernstein, Bradley
    New advances in genomic technology make it possible to address some of the most fundamental questions in biology for the first time. They also highlight a need for new approaches to analyze and model massive amounts of complex data. In this thesis, I present six research projects that illustrate the exciting interaction between high-throughput genomic experiments, new machine learning algorithms, and mathematical modeling. This interdisci- plinary approach gives insights into questions ranging from how variations in the epigenome lead to diseases across human populations to how the slime mold finds the shortest path. The algorithms and models developed here are also of interest to the broader machine learning community, and have applications in other domains such as text modeling.
  • Thumbnail Image
    Publication
    Contrastive Learning Using Spectral Methods
    (Neural Information Processing Systems Foundation, 2013) Zou, James; Hsu, Daniel; Parkes, David; Adams, Ryan Prescott
    In many natural settings, the analysis goal is not to characterize a single data set in isolation, but rather to understand the difference between one set of observations and another. For example, given a background corpus of news articles together with writings of a particular author, one may want a topic model that explains word patterns and themes specific to the author. Another example comes from genomics, in which biological signals may be collected from different regions of a genome, and one wants a model that captures the differential statistics observed in these regions. This paper formalizes this notion of contrastive learning for mixture models, and develops spectral algorithms for inferring mixture components specific to a foreground data set when contrasted with a background data set. The method builds on recent moment-based estimators and tensor decompositions for latent variable models, and has the intuitive feature of using background data statistics to appropriately modify moments estimated from foreground data. A key advantage of the method is that the background data need only be coarsely modeled, which is important when the background is too complex, noisy, or not of interest. The method is demonstrated on applications in contrastive topic modeling and genomic sequence analysis.
  • Thumbnail Image
    Publication
    Strategic Voting Behavior in Doodle Polls
    (Association for Computing Machinery, 2015) Zou, James; Meir, Reshef; Parkes, David
    Finding a common time slot for a group event is a daily conundrum and illustrates key features of group decision-making. It is a complex interplay of individual incentives and group dynamics. A participant would like the final time to be convenient for her, but she is also expected to be cooperative towards other people's preferences. We combine large-scale data analysis with theoretical models from the voting literature to investigate strategic behaviors in event scheduling. We analyze all Doodle polls created in the US from July-September 2011 (over 340,000 polls), consisting of both hidden polls (a user cannot see other responses) and open polls (a user can see all previous responses). By analyzing the differences in behavior in hidden and open polls, we gain unique insights into strategies that people apply in a natural decision-making setting. Responders in open polls are more likely to approve slots that are very popular or very unpopular, but not intermediate slots. We show that this behavior is inconsistent with models that have been proposed in the voting literature, and propose a new model based on combining personal and social utilities to explain the data.
  • Thumbnail Image
    Publication
    Are you Going to Do That? Contingent-Payment Mechanisms to Improve Coordination
    (2015) Ma, Hongyao; Meir, Reshef; Parkes, David; Zou, James
    In this extended abstract, we consider simple coordination problems, such as allocating the right to use a shared sports facility in a way that maximizes its usage, or picking the time of a meeting in a way that maximizes attendance. More generally, an alternative is selected by a mechanism in period zero based on reports from agents. This induces a decision problem facing agents in the next period (e.g., to use a resource, or to attend a meeting.) Outcomes are designated as either good or bad, and the design goal is to maximize the probability of good outcomes. For example, a good outcome may be the resource being used, or having enough people attend a meeting.
  • Thumbnail Image
    Publication
    Tolerable Manipulability in Dynamic Assignment without Money
    (Association for the Advancement of Artificial Intelligence Press, 2010) Zou, James; Gujar, Sujit; Parkes, David
    We study a problem of dynamic allocation without money. Agents have arrivals and departures and strict preferences over items. Strategyproofness requires the use of an arrival-priority serial-dictatorship (APSD) mechanism, which is ex post Pareto efficient but has poor ex ante efficiency as measured through average rank efficiency. We introduce the scoring-rule (SR) mechanism, which biases in favor of allocating items that an agent values above the population consensus. The SR mechanism is not strategyproof but has tolerable manipulability in the sense that: (i) if every agent optimally manipulates, it reduces to APSD, and (ii) it significantly outperforms APSD for rank efficiency when only a fraction of agents are strategic. The performance of SR is also robust to mistakes by agents that manipulate on the basis of inaccurate information about the popularity of items.
  • Thumbnail Image
    Publication
    Priors for Diversity in Generative Latent Variable Models
    (Curran Associates, Inc., 2012) Zou, James; Adams, Ryan Prescott
    Probabilistic latent variable models are one of the cornerstones of machine learning. They offer a convenient and coherent way to specify prior distributions over unobserved structure in data, so that these unknown properties can be inferred via posterior inference. Such models are useful for exploratory analysis and visualization, for building density models of data, and for providing features that can be used for later discriminative tasks. A significant limitation of these models, however, is that draws from the prior are often highly redundant due to i.i.d. assumptions on internal parameters. For example, there is no preference in the prior of a mixture model to make components non-overlapping, or in topic model to ensure that co-occurring words only appear in a small number of topics. In this work, we revisit these independence assumptions for probabilistic latent variable models, replacing the underlying i.i.d. prior with a determinantal point process (DPP). The DPP allows us to specify a preference for diversity in our latent variables using a positive definite kernel function. Using a kernel between probability distributions, we are able to define a DPP on probability measures. We show how to perform MAP inference with DPP priors in latent Dirichlet allocation and in mixture models, leading to better intuition for the latent variable representation and quantitatively improved unsupervised feature extraction, without compromising the generative aspects of the model.
  • Thumbnail Image
    Publication
  • Thumbnail Image
    Publication
    Religion and HIV in Tanzania: influence of religious beliefs on HIV stigma, disclosure, and treatment attitudes
    (BioMed Central, 2009) Zou, James; Yamanaka, Yvonne; John, Muze; Watt, Melissa; Ostermann, Jan; Thielman, Nathan
    Background: Religion shapes everyday beliefs and activities, but few studies have examined its associations with attitudes about HIV. This exploratory study in Tanzania probed associations between religious beliefs and HIV stigma, disclosure, and attitudes toward antiretroviral (ARV) treatment. Methods: A self-administered survey was distributed to a convenience sample of parishioners (n = 438) attending Catholic, Lutheran, and Pentecostal churches in both urban and rural areas. The survey included questions about religious beliefs, opinions about HIV, and knowledge and attitudes about ARVs. Multivariate logistic regression analysis was performed to assess how religion was associated with perceptions about HIV, HIV treatment, and people living with HIV/AIDS. Results: Results indicate that shame-related HIV stigma is strongly associated with religious beliefs such as the belief that HIV is a punishment from God (p < 0.01) or that people living with HIV/AIDS (PLWHA) have not followed the Word of God (p < 0.001). Most participants (84.2%) said that they would disclose their HIV status to their pastor or congregation if they became infected. Although the majority of respondents (80.8%) believed that prayer could cure HIV, almost all (93.7%) said that they would begin ARV treatment if they became HIV-infected. The multivariate analysis found that respondents' hypothetical willingness to begin ARV treatme was not significantly associated with the belief that prayer could cure HIV or with other religious factors. Refusal of ARV treatment was instead correlated with lack of secondary schooling and lack of knowledge about ARVs. Conclusion: The decision to start ARVs hinged primarily on education-level and knowledge about ARVs rather than on religious factors. Research results highlight the influence of religious beliefs on HIV-related stigma and willingness to disclose, and should help to inform HIV-education outreach for religious groups.