Intel Science & Technology Center for Big Data

10/03/2018

After the ISTC for Big Data, what would *you* do for an encore?

Google, Intel and Microsoft have now teamed up with researchers from CSAIL - MIT for an exciting new research project that will push ML and AI deeper into the enterprise. .

Dubbed the “Data Systems and Artificial Intelligence Lab” (DSAIL), the effort will focus on investigating how machine learning can go beyond perception and simple prediction, to improve traditional enterprise applications and data-processing systems.

Headed up by CSAIL professors Sam Madden (pictured) and Tim Kraska (who are co-directors) and 2014 Turing Award winner & CSAIL professor Michael Stonebraker. DSAIL builds on some of the work and code published by the ISTC, with some interesting new themes and researchers.

Read the DSAIL news release here. https://www.csail.mit.edu/news/google-intel-and-microsoft-team-wcsail-new-data-driven-initiative. Follow DSAIL at http://dsail.csail.mit.edu/.

09/01/2017

The Intel Science & Technology Center for Big Data was a five-year, Intel-backed academic-industry research initiative to expedite new Big Data infrastructure technologies. It ran from August 2012 to August 2017. It brought together some of the leading minds working in database and related technology today, from eight top universities and Intel Labs.

Read what the ISTC accomplished in this blog post by ISTC co-directors Michael Stonebraker (2014 Turing Award winner) and Samuel Madden of CSAIL - MIT with Timothy Mattson of Intel Labs.

A full history of the project is available at http://istc-bigdata.org/ #&panel1-2 including links to open-source software produced by the ISTC.

The Big Data ISTC is a research project sponsored by Intel that ran for five years (August 2012- August 2017). This blog post highlights some of the accomplishments and lessons learned during this period.

09/01/2017

This week at VLDB 2017, ISTC PI Arvind and colleagues at CSAIL - MIT presented a new system for data center caching that uses flash memory, the kind of memory used in most smartphones, to make data centers operate more efficiently.

Flash-memory system could reduce power consumption of data center “caches” by 90 percent.

08/30/2017

ISTC PI Timothy Mattson of Intel Labs closes out the VDLB Conference in Munich on Friday,September 1 with this provocatively titled keynote: "Alice in Wonderland, the Red Queen's Race, and Microprocessor Design in a World of Deep Learning."

It's part of the ADMS 2017 Workshop (ADMS = Accelerating Analytics and Data Management Systems Using Modern Processor and Storage Architectures).

08/30/2017

Professor Andy Pavlo and his students at Carnegie Mellon University are working to make sure database management systems (DBMSs) can take advantage of new and evolutionary hardware.

Tomorrow (8/31, 10:55) at the VLDB 2017 Conference in Munich, PhD candidate Joy Arulraj will present the paper “Write-Behind Logging”: how to change a DBMS’s logging and recovery algorithms to use emerging non-volatile memory (NVM). Their new protocol enables the DBMS to recover nearly instantaneously from system failures and reduces the wear-down on NVM devices. This work is in collaboration with Intel Labs and is part of CMU’s extensive work on non-volatile memory databases. Read their blog post about Write-Behind Logging here http://istc-bigdata.org/index.php/write-behind-logging

08/29/2017

Congrats to ISTC technical advisor Pradeep Dubey and Intel researchers Nadathur Satish and Narayanan Sundaram on their paper "Deep Learning at 15PF: Supervised and Semi-Supervised Classification for Scientific Data.. It describes what the authors believe to be the most scalable deep-learning implementation in the world currently, and runs on Intel Xeon Phi technology.

The work involved a collaboration between Intel, the National Energy Research Scientific Computing Center, and Stanford.

Read the article on HPCWire

A collaborative effort between Intel, NERSC and Stanford has delivered the first 15-petaflops deep learning software running on HPC platforms and is, accor

08/29/2017

ISTC PI Professor Bill Howe of University of Washington is moderating a panel on "The Information War: Fake News, Privacy and Big Data" on September 18 in Seattle.

This event is hosted by University of Washington Master of Science in Data Science in partnership with UW Professional and Continuing Education and the eScience Instutute (which he heads).

An information war is underway. Politicians, advertisers and even scientists are using the internet (and your own data) to commandeer your attention and influence your actions. They know who you are and what you’ll respond to. Some, we’ve seen, are not above promoting half-truths, making predictions...

08/28/2017

The research papers & demo sessions kick off at the VLDB 2017 Conference tomorrow, August 29, in Munich. Check out some of the talks and demos being given by ISTC Researchers, starting with:

Intel's Stavros Papadopoulos and colleagues present “The TileDB Array Data Storage Manager,” a novel storage system that captures cases where the data are naturally modeled as multi-dimensional arrays, which typically arise in scientific applications. TileDB has been used in genomics as the storage layer of GenomicsDB, which is maintained by the Intel Health and Life Sciences group and is used in the Broad Institute‘s GATK 4.0 software. Both TileDB and GenomicsDB are available in open source. Wed, 8/30 13:30 .Read more about TileDB in this blog post. http://istc-bigdata.org/index.php/tiledb-and-genomicsdb-now-available-in-open-source/

Researchers in Brown University's Data Management Research Group will present “Revisiting Reuse for Approximate Query Processing.” They’ve developed a formulation for AQP that maximizes result reuse in order to improve interactivity for visual data exploration tasks. Their AQP formulation can provide low-error approximate results at interactive speeds, even for queries over rare sub-populations. Wed, 8/30 10:30

08/28/2017

Secure sharing of private and regulated data for open science, business and other analysis could lead to advancements that benefit us all. But protecting data is a challenge. For example: many hospitals are interested in pooling their medical records for research, but none may disclose arbitrary patient records to researchers or other healthcare providers, write ISTC PI Jennie Rogers and her colleagues from Northwestern University.

The Northwestern team has proposed a possible solution: the Private Data Network (PDN), a federated database for querying over the collective data of mutually distrustful parties. They will be presenting their published paper at the Very Large Databases Conference (VLDB 2017) in Munich tomorrow, Tuesday, at 13:30.

You can download their paper, “SMCQL: Secure Query Processing for Private Data Networks,” here.

08/27/2017

How data gathering and use can change the world: Prediction, Prevention, Protection. Our research partner Intel Labs puts a premium on research that can affect billions of people around the world.

08/25/2017

ISTC researchers at Portland State University and University of Washington have been on the cutting edge of the Smart Cities movement in the Northwest, innovating in urban analytics and advocating for open data and responsible algorithmic decision making.

"Bill Howe, associate director of the eScience Institute at the UW, said the region can lead with policies that ensure 'responsible algorithmic decision making,' or data-driven models that take into account human observation. He called it a “huge opportunity.”

Read PSU Professor (and Smart Cities activist) Kristin Tufte's most recent blog post on how urban analytics can enable Smart Cities through improved transportation.

http://istc-bigdata.org/index.php/urban-analytics-for-smart-cities-connecting-data-to-people/

New technology is helping local government create “smarter” cities in a variety of ways, from adaptive traffic lights to open data platforms to advanced utility meters. But with innovation comes…

08/25/2017

At VLDB 2017 on Monday, August 28, the intriguingly titled FADS Workshop (Failed Aspirations in Database Systems) looks back at promising ideas in database systems technology that, for whatever reason, just didn't go - and possible lessons for computer scientists & researchers.

Impressive line-up, including Phil Bernstein of Microsoft Research and Professor Margo Seltzer of Harvard University (Berkeley DB/ Sleepycat Software). ISTC PI Andy Pavlo of Carnegie Mellon University is workshop chair.

Here's the full program line-up (10:30 - 17:00): http://fads.ws/program/

Failed Aspirations in Database Systems --VLDB Workshop (Aug 28th, 2017)