AT&T Labs will be hosting the NYC WiMLDS community for an Intro to PyTorch event with Dr. Michela Paganini from Facebook AI Research! We would also like to thank Facebook for sponsoring this event. Links to Google Colab notebooks will be provided at the event. She joined Facebook in 2018 after earning her PhD in particle physics from Yale University under the supervision of Paul Tipton. Her work focuses on the empirical characterization of neural network dynamics using tools from theoretical and experimental physics.
This is the conference, and here's my talk (will do Google hangout, just as with my recent talks in Bern, Strasbourg, etc): Through a series of examples, we consider problems with classical hypothesis testing, whether performed using classical p-values or confidence intervals, Bayes factors, or Bayesian inference using noninformative priors. We locate the problem not in the use of any particular statistical method but rather with larger problems of deterministic thinking and a misguided version of Popperianism in which the rejection of a straw-man null hypothesis is taken as confirmation of a preferred alternative. We suggest solutions involving multilevel modeling and informative Bayesian inference. The post Hypothesis Testing is a Bad Idea (my talk at Warwick, England, 2:30pm Thurs 15 Sept) appeared first on Statistical Modeling, Causal Inference, and Social Science. The post Hypothesis Testing is a Bad Idea (my talk at Warwick, England, 2:30pm Thurs 15 Sept) appeared first on All About Statistics.
At Convoy, we've built a digital marketplace to connect truck driving businesses to our customer's shipments as well as the infrastructure to collect comprehensive data on each transaction and user engagement with our technology. Our data have a complex underlying network structure and unique supply/demand equilibrium constraints. In this talk, I will review my exploratory work to create a synthetic dataset of bidding behavior that exhibits interference and evaluate analysis of a 2-stage randomization design experiment. My goal is to provide insight into the intricacies of a digital marketplace, using Convoy as a specific example, and difficulties with applying solutions developed for the general interference setting.
WiMLDS's mission is to support and promote women and gender minorities who are practicing, studying or are interested in the fields of machine learning and data science. We create opportunities for members to engage in technical and professional conversations in a positive, supportive environment by hosting talks by women and gender minority individuals working in data science or machine learning. Events include technical workshops, networking events and hackathons. We are inclusive to anyone who supports our cause regardless of gender identity or technical background.
Someone asked me about the distinction between bias and noise and I sent him some links. Here's a recent paper on election polling where we try to be explicit about what is bias and what is variance: And here are some other things I've written on the topic: – The bias-variance tradeoff – Everyone's trading bias for variance at some point, it's just done at different places in the analyses – There's No Such Thing As Unbiased Estimation. Finally, here's the sense in which variance and bias can't quite be distinguished: – An error term can be mathematically expressed as "variance" but if it only happens once or twice, it functions as "bias" in your experiment. An experimental protocol could be positively biased one day and negatively biased another day or in another scenario. The post Noise and bias appeared first on Statistical Modeling, Causal Inference, and Social Science.