Alex Smola
Adventures in Data Land: Alex Smola's blog on machine learning, AI, and systems.
- Posts
- 48 from 15 Jan 2010 to 9 Aug 2026
- Posting
- monthly (15 in the last year)
- Kind
- personal blog
- Feed
- http://alex.smola.org/blog.xml
2026 15
- AI for humans
- Losing the thread
- The autopsy: porting MXNet to Blackwell and Apple Silicon
- Raising MXNet from the Attic
- Efficiency in LLMs
- Fast projections require a PHD
- Random projections, three ways
- Digital humans
- The effective sample size
- ProactBench
- What your assistant didn’t say
- Beating \(2^n - 1\) for InfiniBand
- Fair representations are pancakes
- You don’t need all the benchmarks
- The Pokémon Theorem
2019 1
2016 1
2014 1
2013 1
2012 8
- 100 Terabytes, 5 Billion Documents, 10 Billion Parameters, 1 Billion Inserts/s
- Beware the bandwidth gap - speeding up optimization
- The Weisfeiler-Leman algorithm and estimation on graphs
- In defense of keeping data private
- MLSS Purdue
- Random numbers in constant storage
- May 15, 2012
- The Neal Kernel and Random Kitchen Sinks
2011 18
- Big Learning: Algorithms, Systems, and Tools for Learning at Scale
- Introduction to Graphical Models
- Distributed synchronization with the distributed star
- Speeding up Latent Dirichlet Allocation
- Bloom Filters
- Jul 25, 2011
- Jul 20, 2011
- Jul 15, 2011
- Jul 10, 2011
- Jul 5, 2011
- Jul 1, 2011
- Jun 30, 2011
- Jun 25, 2011
- Jun 20, 2011
- May 1, 2011
- Apr 1, 2011
- Mar 1, 2011
- Jan 1, 2011