AMPLab – UC Berkeley

AMPLab - UC Berkeley · amplab.cs.berkeley.edu · American English

Algorithms, Machines and People Lab

rss2 wordpress content atom dc sy

Open the feed

https://amplab.cs.berkeley.edu/feed/

Last post: Apr 26, 2017
Posts in 24 h · 7 days · 30 days: 0 · 0 · 0
Our last check: Answering
Served from: United States
Site title: AMPLab - UC Berkeley
Text score at discovery: 6,795
Format: rss2
Features in the feed: content, atom, dc, sy
Community: wordpress

Posts

What our queue read from this feed. Open one to read it here, or go to the site that published it.

Making cities safer: data collection for Vision Zero
Apr 26, 2017 · original
A critical part of enabling cities to implement their Vision Zero policies – the goal of the current National Transportation Data Challenge – is to be able to generate open, multi-modal travel experience data. While existing datasets use police and hospital reports to provide a comprehensive picture of fatalities and life altering injuries, by their nature, they are sparse and resist use for prediction and prioritization. Further, changes to infrastructure to support Vision Zero policies frequently require balancing competing needs from different constituencies – protected bike lanes, dedicated signals and expanded sidewalks all raise concerns that automobile traffic will be severely impacted. A timeline of the El Monte/Marich intersection in Mountain View, from 2014 to 2017 provides an opportunity to put some of these challenges into context. since there is no standard way to report nea
Data Hackathons Workshop: Early Career Funding Available
Sep 1, 2016 · original
The West Big Data Innovation Hub (WBDIH) is excited to host a Data Hackathons: Lessons Learned and Best Practices Workshop on September 15 as part of the historic first International Data Week in Denver. One of four hubs recently launched with funding from the National Science Foundation and leadership from UC Berkeley, University of Washington, and the San Diego Supercomputer Center, the WBDIH builds and strengthens partnerships across industry, academia, nonprofits, and government to address societal challenges. As mentioned at the WBDIH All Hands Meeting held at UC Berkeley this spring, the workshop will convene hackathon organizers, sponsors, and other stakeholders to share insights about the design, implementation, scalability, and impact of data-focused hackathons. Data hackathon case studies mentioned will cover the WBDIH thematic areas including Metro Data Science, Natural Resour
CACM Article on Randomized Linear Algebra
Jun 2, 2016 · original
Each month the Communications of the ACM publishes an invited “Review Article” paper chosen from across the field of Computer Science. These papers are intended to describe new developments of broad significance to the computing field, offer a high-level perspective on a technical area, and highlight unresolved questions and future directions. The June 2016 issue of CACM contains a paper by AMPLab researcher Michael Mahoney and his colleague Petros Driness (of RPI, soon to be Purdue). The paper, “RandNLA: Randomized Numerical Linear Algebra,” describes how randomization offers new benefits for large-scale linear algebra computations such as those that underlie a lot of the machine learning that is developed in the AMPLab and elsewhere. Randomized Numerical Linear Algebra (RandNLA), a.k.a., Randomized Linear Algebra (RLA), is an interdisciplinary research area that exploits randomization
Scientific Matrix Factorizations In Spark at Scale
Jun 2, 2016 · original
The canonical example of matrix decompositions, the Principal Components Analysis (PCA), is ubiquitous, with applications in many scientific fields including neuroscience, genomics, climatology, and economics. Increasingly, the data sets available to scientists are in range of hundreds of gigabytes or terabytes, and their analyses are bottle-necked by the computation of the PCA or related low-rank matrix decompositions like the Non-Negative Matrix factorization (NMF). The sheer size of these data sets necessitates distributed analyses. Spark is a natural candidate for implementation of these analyses. Together with my collaborators at Berkeley: Aditya Devarakonda , Michael Mahoney , James Demmel , and with teams at Cray, Inc. and NERSC’s Data Analytic Services group , I have been investigating the performance of Spark at computing scientific matrix decompositions. We used MLlib and ml-ma
Technical Preview of Apache Spark 2.0: Easier, Faster, and Smarter
May 19, 2016 · original
This is a guest blog post originally published on the Databricks blog . For the past few months, we have been busy working on the next major release of the big data open source software we love: Apache Spark 2.0. Since Spark 1.0 came out two years ago, we have heard praises and complaints. Spark 2.0 builds on what we have learned in the past two years, doubling down on what users love and improving on what users lament. While this blog summarizes the three major thrusts and themes—easier, faster, and smarter—that comprise Spark 2.0, the themes highlighted here deserve deep-dive discussions that we will follow up with in-depth blogs in the next few weeks. Prior to the general release, a technical preview of Apache Spark 2.0 is available on Databricks . This preview package is built using the upstream branch-2.0. Using the preview package is as simple as selecting the “2.0 (branch preview)
AMPLab postdoc Julian Shun wins the ACM Doctoral Dissertation Award
May 11, 2016 · original
I am very pleased to announce that Julian Shun has been awarded the ACM’s doctoral dissertation award for his 2015 CMU doctoral thesis “ Shared-Memory Parallelism Can Be Simple, Fast, and Scalable ” which also won that year’s CMU SCS distinguished dissertation award. Julian currently works with me as a postdoc both in the Department of Statistics and in the AMP Lab in the EECS Department and is supported by a Miller Fellowship. His research focuses on fundamental theoretical and practical questions at the interface between computer science and statistics for large-scale data analysis. He is particularly interested in all aspects of parallel computing, especially parallel graph processing frameworks, algorithms, data structures and tools for deterministic parallel programming; and he has developed Ligra , a lightweight graph processing framework for shared memory. More details can be foun
Unexpected continuous location tracking/energy change in android?
Apr 15, 2016 · original
So this is really weird, but I have found what seems to be unexpected continuous location tracking that is causing noticeable battery drain on Android 6.0. Right now, it’s looking like a change in an automatically updated component, so it is probably due to a closed source service or app. So this is in the style of the work from Vern Paxson’s group on characterizing the observed behavior of third party software. Has anybody else running Android 6.0 noticed a particularly large increase in power drain, with the GPS icon displayed continuously? I will be running additional tests in the coming days, but wanted to report the unusual behavior and see if other researchers have noticed it as well, or want to investigate it while it lasts. Background I’ve been doing power profiling of power drain under various regimes as part of understanding the power/accuracy tradeoffs for my travel pattern tr
Strata Keynote: “What’s Next for the Berkeley Data Analytics Stack”
Apr 1, 2016 · original
The Strata+Hadoop World big data industry conference was held in San Jose this week and as usual, AMPLab and Berkeley were well-represented among the presentations, while uniquely among university projects, AMPLab alumni and AMPLab-developed software were prominent throughout the conference. I got the opportunity to give an update on the BDAS stack in a short Keynote talk to the (I’m guessing) 2500 or so people who got up early enough to attend the morning session on Thursday. Scheduled between a talk on brain surgery and a stand up improv routine by Paula Poundstone (of NPR’s “ Wait Wait, Don’t Tell Me “) and after a nice introduction from long-time AMPLab friend and supporter Ben Lorica , I gave a quick over view of the AMPLab and BDAS, and then focused on four of our ongoing projects: Succinct , Velox , KeystoneML , and AMPCrowd/SampleClean . O’Reilly, the sponsor of Strata, has put v
40-Year Goodbye: A Last Lecture and Symposium
Mar 29, 2016 · original
Hoping to start a new tradition, I’m giving a Last Lecture on Friday May 6, 2016 at 4PM , which is shortly before I retire . (The premise of these is that if this were the last public lecture you would give, what would you say.) My title is “ How to Be a Bad Professor ” ( abstract ), and it is followed by a reception. There will also be a one-day symposium on Saturday May 7 ( agenda ) with talks by colleagues and former students on the future of topics associated with my 40 years at UC Berkeley, such as the microprocessors, storage, cloud computing, data science, and machine learning. Both events will be held at the International House at UC Berkeley. Please sign up to the event of interest , and feel free to invite other interested parties. Dave P.S. Those who sign up before April 7 for the symposium will get a commemorative book with perspectives by Stanford President John Hennessy, Na
What’s new in KeystoneML
Mar 29, 2016 · original
At the AMPLab, we are constantly looking for ways to improve the performance and user experience of large scale advanced analytics. We frequently make the fruits of our research available as open source software. Last week, we released version 0.3 of KeystoneML , a project we’ve blogged about in the past. KeystoneML is designed to simplify the construction of large scale end-to-end machine learning pipelines. For this development cycle, we focused our efforts on pipeline optimization including new features to automatically materialize intermediate reused state and cost-based selection of logical pipeline operators. In this post, we’ll recap how users can describe machine learning applications using high-level operators with KeystoneML. Then, we’ll discuss how the new features in the latest release accelerate the training of these applications . In KeystoneML, users describe their machine

Discovered by the rss-feed-index crawler, which checks each feed at most once a month.

Same record as JSON: https://api.agentalog.com/api/feeds/fd_amplab_cs_berkeley_edu_c59c7d9cf6349a72. More from this site: amplab.cs.berkeley.edu in the Feeds tab.