Presentations

4/13/2016 - Marie-Abele Bind (Harvard) - Valid and informative p-values from big data, illustrated in an epigenomic cross-over experiment, at CGIS Knafel K354, Wednesday, April 13, 2016

Title: Valid and informative p-values from big data, illustrated in an epigenomic cross-over experiment

Abstract: A common issue that arises with current analyses of epigenomic data is the repeated use of statistical tests. For example, consider 17 people in a randomized experiment measuring the results of exposure to two treatment conditions (e.g., clean air and ozone), with measurements at 484,531 epigenome locations, where the aim is to find the locations with an epigenetic effect (i.e., of clean air versus ozone). Here, we describe the use of...

Read more about 4/13/2016 - Marie-Abele Bind (Harvard) - Valid and informative p-values from big data, illustrated in an epigenomic cross-over experiment
4/6/2016 - Ethan Fosse & Chris Winship (Harvard) - Bounding Analyses of Age-Period-Cohort Models, at CGIS Knafel K354, Wednesday, April 6, 2016

Title: Bounding Analyses of Age-Period-Cohort Models

Abstract: 

For at least 80 years researchers in a wide variety of fields have sought to uniquely identify age, period, and cohort (APC) effects, even though an infinite number of solutions exist due to perfect linear dependency. In this paper we introduce a new approach for identifying APC effects based on bounding feasible regions of the parameter space. Depending on the location of the solution line in the parameter space, minimal constraints on the direction and magnitude of the linear trends...

Read more about 4/6/2016 - Ethan Fosse & Chris Winship (Harvard) - Bounding Analyses of Age-Period-Cohort Models
3/30/2016 - Nan Laird (Harvard) - Multivariate Problems in the Genetic Analysis of Complex Disease, at CGIS Knafel K354, Wednesday, March 30, 2016

Title: Multivariate Problems in the Genetic Analysis of Complex Disease

Abstract: Complex diseases have multiple underlying contributing factors, both genetic and environmental.  In addition, the disease syndrome is often characterized by numerous clinical traits that may be analyzed for association with genes along with the disease status.  Genome Wide Association Analysis (GWAS) has been highly successful in identifying some genetic loci associated with many disease syndromes and/or selected traits.  The purpose of the analysis of multiple traits may be to...

Read more about 3/30/2016 - Nan Laird (Harvard) - Multivariate Problems in the Genetic Analysis of Complex Disease
3/23/2016- Laura Balzer- Targeted Learning in the SEARCH trial and HIV prevention in East Africa Wednesday, March 23, 2016

Title: Targeted Learning in the SEARCH trial and HIV prevention in East Africa

Abstract: 

Evaluation of community-based interventions presents significant methodological challenges. In this talk, we describe the design and analysis of the SEARCH trial, an ongoing community randomized trial to evaluate the impact of early HIV diagnosis and immediate treatment with streamlined care in rural East Africa. We focus on 3 choices to optimize study power: adaptive pair-matching over complete randomization,...

Read more about 3/23/2016- Laura Balzer- Targeted Learning in the SEARCH trial and HIV prevention in East Africa
3/9/2016- Stefanie Jegelka (MIT)- Algorithms and new applications for determinantal point processes Wednesday, March 9, 2016

Title: Algorithms and new applications for determinantal point processes

Abstract:  Many real-world inference problems are, at their core, subset selection problems. Probabilistic models for such scenarios rely on having distributions over discrete sets that are sufficiently accurate yet computationally efficient to work with. We focus on sub-families of such distributions whose special mathematical properties are the basis for fast algorithms. As a specific example, Determinantal Point Processes (DPPs) have recently become popular in...

Read more about 3/9/2016- Stefanie Jegelka (MIT)- Algorithms and new applications for determinantal point processes
3/2/2016- Finale Doshi-Velez (Harvard)- Cross-Corpora Learning of Trajectories in Autism Spectrum Disorders Wednesday, March 2, 2016

Title: Cross-Corpora Learning of Trajectories in Autism Spectrum Disorders

Abstract: 

Patients with developmental disorders, such as autism spectrum disorder (ASD), present with symptoms that change with time even if the named diagnosis remains fixed.  For example, a child may have delayed speech as a toddler and difficulty reading in elementary school.  Characterizing these trajectories is important for early treatment.  However, deriving these trajectories from observational sources is challenging:...

Read more about 3/2/2016- Finale Doshi-Velez (Harvard)- Cross-Corpora Learning of Trajectories in Autism Spectrum Disorders
2/24/2016- Jessie Myers Franklin (Harvard & Brigham Women's)- Comparing marginal estimators of propensity-adjusted treatment effects in studies with few observed outcome events Wednesday, February 24, 2016

Title: Comparing marginal estimators of propensity-adjusted treatment effects in studies with few observed outcome events

Abstract:  Nonrandomized studies of treatments from electronic healthcare databases are critical for producing the evidence necessary to making informed treatment decisions, but often rely on comparing rates of events observed in a small number of patients. In addition, a typical study constructed from an electronic healthcare database, for example, administrative claims data, requires adjustment for many,...

Read more about 2/24/2016- Jessie Myers Franklin (Harvard & Brigham Women's)- Comparing marginal estimators of propensity-adjusted treatment effects in studies with few observed outcome events
2/17/2016- Jann Spiess (Harvard)- Robust Post-Matching Inference Wednesday, February 17, 2016

Title: Robust Post-Matching Inference

Abstract: 

Nearest-neighbor matching (Cochran, 1953; Rubin, 1973) is a popular nonparametric tool to create balance between treatment and control groups in non-experimental data. As a preprocessing step for regression analysis, it reduces the dependence on parametric modeling assumptions (Ho et al., 2007). In this paper, we show how to obtain valid standard error estimates for linear regression after nearest-neighbor matching without replacement. We show that standard error estimates...

Read more about 2/17/2016- Jann Spiess (Harvard)- Robust Post-Matching Inference
2/10/2016- Hanna Wallach (Microsoft Research) Modeling Topic-Partitioned Network Structure Wednesday, February 10, 2016

Title: Modeling Topic-Partitioned Network Structure

Abstract: 

In this talk, I will discuss two projects centered around modeling
topic-partitioned network structure. The first focuses on obtaining
and analyzing local government email corpora. I will describe a field
experiment that we conducted to investigate whether governments'
compliance with public records requests is influenced by the knowledge
that their peers have already complied. I will then talk about
studying local government organizations...

Read more about 2/10/2016- Hanna Wallach (Microsoft Research) Modeling Topic-Partitioned Network Structure
2/3/2016- Nicole Immorlica (Microsoft Research)- The Degree of Segregation in Social Networks Wednesday, February 3, 2016

Abstract: In 1969, economist Thomas Schelling introduced a landmark model of racial segregation in which individuals choose residences based on the racial composition of the corresponding neighborhoods.  Simple simulations of Schelling's model suggest this local behavior can cause segregation even for racially tolerant individuals.  In this talk, we provide rigorous analyses of the degree of segregation in Schelling's model on one-dimensional and two-dimensional lattices.  We see that if agents refuse to live in neighborhood in which their...

Read more about 2/3/2016- Nicole Immorlica (Microsoft Research)- The Degree of Segregation in Social Networks

Pages