TALKS AND DEMOS (FROM MODERN TO ANCIENT)

A talk on using maximum entropy to estimate the prevalence of multiple chronic diseases.
Our VLDB 2020 demo talk on the sPaQLTools interface for stochastic package queries. ** VLDB 2020 Best Demonstration Award**
Our VLDB 2020 demo talk on the SuDocu prototype system for automatic document summarization. ** VLDB 2020 Best Demonstration Runner-up**
Our SIGMOD 2020 talk on stochastic package queries in probabilistic databases. (Starts at approx. 33 minutes into the session.)
An interview by UMass Amherst INFORMS Student Chapter, 2019
A brief overview talk on research challenges in approximate query processing presented at EDBT 2018.
Link to open-source SimSQL code and tutorials. (Not actually a demo, since SimSQL is a fully functioning system.)
A talk and poster for the Foresight visualization recommender prototype (KDD IDEA 2017 and VLDB 2017.)
A 2016 talk on compressed linear algebra for machine learning (VLDB 2016).
A 2015 talk about Splash: Smarter Planet Platform for Analysis and Simulation of Health. (video here).
A 2014 talk on model-data ecosystems (the interface between simulation and data science).
A talk on indexing in Hadoop (EDBT 2013).
A brief overview talk on Ricardo and DSGD for matrix factorization (XLDB 2011).
A talk on risk management as a driver of database technology for managing uncertain data (keynote talk, 2010 VLDB Workshop on Management of Uncertain Data).
A 2010 talk on the Monte Carlo Database System (MCDB) for stochastic analytics close to the data.
A talk on transience and recurrence in irreducible finite-state stochastic systems (ETH Mathematics Department Seminar, 2009).
A talk on maintaining a Bernoulli sample over an evolving multiset (PODS 2007).
SIGMOD 2007 "Test of Time" talk about online aggregation.
A talk about entity-resolution-aware query answering for BI.
A 2007 tutorial on simulation of stochastic Petri nets (ICATPN '07).
A talk on maintaining a bi-level Bernoulli sampling scheme for relational databases (SIGMOD 2004).
Talks on discovering correlations and soft functional dependencies in a database, via sampling (SIGMOD 2004) and query feedback (VLDB 2007).
A 2003 talk on how to speed up DB2 using sampling, as well as an extended version of a column on sampling in DB2 that appeared in the IDUG Solutions Journal in 2003.
A talk on a "bump hunting" method for discovering soft algebraic constraints. Led to "best of breed" TPC-H performance for DB2 LUW (VLDB 2003).
A 2002 talk on estimation methods for discrete-event systems.
A talk on how to do some simple statistical analysis in DB2, with examples. Some of this material can also be found in an IBM Redbook on this topic.

(Home)