HNHacker News
TopNewBestAskShowJobs

banjo_milkman

419 karma · joined October 1, 2015

submissionscomments
banjo_milkman··on Ask HN: What Neural Networks/Deep Learning Books Should I Read?
1. The Goodfellow book is the obvious one. Another option is 'Machine Learning: A probabilistic perspective' by Kevin Murphy, which does have exercises but has less material on CNNs IIRC and is slightly older. Also Bishop 'Pattern Recognition and Machine Learning' is a classic reference but even older. 2. The Pedro Domingos book the Master Algorithm tries to address this itch / may be less philosophical that you want. It is >okay< but I don't think people love it.
banjo_milkman··on Graph Processing on FPGAs: Taxonomy, Survey, Challenges
Many people think FPGAs have the advantage over GPUs in machine learning inference. See the Bing FPGA stuff. Mostly due to lower power operations, but also a higher level of customization e.g. optimized 1 bit math if you want it, leading to potentially higher speed or larger networks. The memory architecture of FPGAs can also offer advantages - e.g. bigger memories.

The GPU instruction set is limited for some applications because of SIMD. If the GPU threads aren't doing exactly the same thing then they stall. The FPGA can have parallel processes going on that are more differentiated. e.g. HMMs work better on FPGA than GPU.

Anything with low latency feedback (e.g. LSTMs) probably works better on FPGA.

banjo_milkman··on Graph Processing on FPGAs: Taxonomy, Survey, Challenges
Not really, but I suppose it depends on the application requirements. It typically takes under a second which is much faster than taping out another chip ;-) but could be problematic for real-time systems in some cases. It depends upon the size of the FPGA/bus speed etc.
banjo_milkman··on Graph Processing on FPGAs: Taxonomy, Survey, Challenges
Back when chip costs were reasonable, custom chips would always beat FPGAs and were an easy option for many applications but I think things have changed somewhat in the last few years with the decline of Moore's law. Now there are fewer and fewer companies that can afford to develop custom chips at leading technology nodes. Some estimates are it costs $100m to get a custom chip into product at 7nm. The volumes just aren't there for many applications to justify a custom chip, it doesn't make sense unless the market is huge and the algorithms are static. FPGAs work well in these cases. Have you seen Microsoft Brainwave? https://www.microsoft.com/en-us/research/project/project-bra...
banjo_milkman··on Ask HN: Biographies recommendations – both Westerner and Non-westerner
The Man Who Knew Infinity: A Life of the Genius Ramanujan by Robert Kanigel ; it's impossible to understand how he figured out the amazing things he did. His story is a tragedy.

Locke: A biography by Roger Woolhouse ; the father of classical liberalism in a time of upheaval in England, interactions with the king, banishment to Holland, promoted religious toleration

The Double Helix: James Watson; scientific discovery in 1950s Cambridge, the thrill of the race to be first, controversy with Rosalind Franklin

The Caro books on Lyndon Johnson; insight into the ugly guts of politics, what it takes to win & the course of mid-20th century American history

The Mark Levison book on the Beatles is amazing - the only volume he has finished takes ~700 pages to get to the release of their first single. Arguably the most important cultural phenomenon of the 20th century.

Andrew Hodge's biography of Alan Turing

banjo_milkman··on Ask HN: Biographies recommendations – both Westerner and Non-westerner
I really enjoyed the Man who knew Infinity by Kanigel. Seemed to capture the flavor of Cambridge at the time. It's a movie too, but I haven't seen it.
banjo_milkman··on Ask HN: Favorite nonfiction books of 2018?
Books I liked in 2018:

Crashed by Adam Tooze ; history of the financial crisis, goes into more detail than most of the others

The Future of Capitalism by Paul Collier; lots of ideas on how to improve our situation, most of them are good, UK focused

Bad Blood by John Carreyrou; if you've ever been in a startup you'll recognize bits of this story, but it quickly gets out of control in novel ways. Astonishing story.

What Money Can't Buy: The Moral Limits of Markets by Michel Sandel

The Attention Merchants by Tim Wu

Who we are and how we got here by David Reich: ancient DNA and human history

The Book of Why by Judea Pearl; liked it but I need to reread this one a few more times to comprehend completely or go to his textbooks

Empire of Cotton: Sven Birckets; a history of the first global technology including how it made the UK & USA rich

The Away Game: The Epic Search for Soccer's Next Superstar by Sebastian Abbot

banjo_milkman··on AWS EC2 FPGA Hardware and Software Development Kit
Another application is accelerating DNA sequencing, the first FPGA-based cloud service on AWS?: http://www.edicogenome.com/applications/dragen-on-amazon-web... https://www.nextplatform.com/2016/12/07/configuring-future-f... http://www.edicogenome.com/news/dragenblog/
← PreviousPage 2 of 2