Radim

42 ms·

391.
by Radim 9y ago
For people wondering what Sumy is: Our students compared and benchmarked Sumy against a bunch of other popular summarization techniques, https://rare-technologies.com/text-summarization-in-python-e...
392.
Toward a universal decoder of linguistic meaning from brain activation(nature.com)
2 points by Radim 9y ago|0 comments
393.
by Radim 9y ago
Let dead things go in peace. Pity Nature doesn't have a Wayback Machine: http://archive.org/web/
394.
A Hundred Authors Against Einstein(weeklysciencequiz.blogspot.com)
2 points by Radim 9y ago|0 comments
395.
by Radim 9y ago
That opening gave me a pause too. Doesn't Pandas target only structured (tabular) in-core datasets? (Unless you use something like Blaze / Dask / Ray on top.) Does anyone really work with Pandas on "100's of terabyt
396.
by Radim 9y ago
I appreciate the offer, but I'm positive you can spot the issues yourself. Gensim has many warts for sure, but a benchmark that gives a green tick to itself in "Simple Pythonic interface" and a red check mark to Gensim, while
397.
by Radim 9y ago
Interesting effort :-) Unfortunately, your comparison table there is somewhere between misleading and insulting. Almost all of the "unique" features listed there are in fact a standard part of Gensim. Fast approximative queries (u
398.
by Radim 9y ago
You might be interested to know that Annoy is also integrated into Gensim, which allows you to train, use and query word embeddings on your own data. Gensim implements fast Doc2Vec and FastText too, which are a bit newer embedding technique
399.
by Radim 9y ago
For people interested in path finding, there was some fascinating research done in the 90s: http://faculty.nps.edu/ncrowe/snell2.htm They used Snell's law to design pathfinding directly on region polygons, based o
400.
This New AI System Can See What You Are Thinking(interestingengineering.com)
2 points by Radim 9y ago|0 comments
401.
by Radim 9y ago
Indeed, and not only yours. The same issue crops up in B2B due diligence and risk management, irrespective of GDPR. The automated inventory analysis tools, like ours and others, are only just becoming useful thanks to ML. The previous gener
402.
by Radim 9y ago
Funnily, one of the common fears our clients ( https://gdpr-tools.eu ) have with regards to GDPR is not about the general public. It comes from disgruntled employees ratting on the company. Employees know best where personal data
403.
by Radim 9y ago
> What's the minimum amount of data? Who decides that? Is it dependent on context? I'd hope so! Elizabeth Denham, UK's information commissioner in charge of data protection enforcement, had this to say: "Having lar
404.
by Radim 9y ago
Of course—if you don't store personal data (trivially). In fact, encryption (security) is mostly orthogonal to how you track and handle personal and sensitive data (privacy protection). You could encrypt everything and still be wildly
405.
by Radim 9y ago
Preprint of the article on BioRxiv (PDF): https://www.biorxiv.org/content/early/2017/12/30/240317
406.
This New AI System Can See What You Are Thinking(interestingengineering.com)
1 points by Radim 9y ago|1 comments
407.
Prehistoric Cave Paintings, Symbolic Thinking, and the Emergence of Language(frontiersin.org)
1 points by Radim 9y ago|0 comments
408.
Prehistoric Cave Paintings, Symbolic Thinking, and the Emergence of Language(frontiersin.org)
1 points by Radim 9y ago|0 comments
409.
by Radim 9y ago
Case in point, collocation detection over very large corpora: https://github.com/RaRe-Technologies/bounter#example-on-the-... Reduces RAM requirements some 31x times when automatically detecting common multiword expres
410.
by Radim 9y ago
There's a number of tools that do entity linking (that's the phrase you're looking for), including some open source ones. We've evaluated and used quite a few over the years: there's Dexter [0] by Diego Ceccarelli
411.
A Different Take on Sexism in Science(blogs.scientificamerican.com)
2 points by Radim 9y ago|0 comments
412.
Ask HN: How to exclude mass media from search results?
2 points by Radim 9y ago|0 comments
413.
by Radim 9y ago
Honest question: I'm from eastern Europe, trying to wrap my head around this US controversy. I fully respect Google's right for at-will termination, just trying to figure out whether Damore sinned by going against "world-as-i
414.
by Radim 9y ago
Scikit-learn. The guys have done, and are continuing to do, an amazing job and the output is completely free. There is some funding involved, but the ultimate value/cost ratio is ridiculously high. "How Open Source is Really Maint
415.
by Radim 9y ago
Actually, the law is defined quite broadly, not restricting itself to "name, IP, email". Have you considered how a combination of innocuous data points, such as "browser + city + top 3 popular sites" can make a person un
416.
by Radim 9y ago
I see the "spirit of the law" and "intent" being thrown around here. Elizabeth Denham, UK's information commissioner in charge of data protection enforcement, had this to say: "Having larger fines is useful bu
417.
The crimes against dopamine(medium.com)
1 points by Radim 9y ago|0 comments
418.
by Radim 9y ago
How many have been awarded to me? There lies the ultimate discrimination resolution—at the level of the individual.
419.
by Radim 9y ago
Various approaches based on LSH (locality sensitive hashing) and random projections. But "the best" really depends on your data profile, as the generic problem formulation, for completely generic vectors, is too… generic. See also
420.
by Radim 9y ago
Slightly OT, isn't that Monty Python movie ageless? In fact, like good wine, it only gets better with time. I gave The Life of Brian another view earlier this week (it's on Youtube!) and was awed by its prescience.

More ›