alew1

14 ms·

31.
by alew1 7y ago
Good question. Disclaimer: I’m in the lab that made Gen & was on the paper, so not impartial :) Anglican is implemented in Clojure, and can be extended (by writing new Clojure code) to support new general-purpose inference engines. Crea
32.
by alew1 8y ago
This is really cool! And awesome that (anecdotally) it seems to have been an effective tool for teaching high school students about logic and proof. Excited to dig in and learn more at some point about how this relates to formal proof assis
33.
by alew1 8y ago
Hmm, I'm not sure I see the difference. Why is it not "symbolic"? The symbols that construct the neural network are what encodes translation invariance -- not some vector of reals.
34.
by alew1 8y ago
As I was reading, I kept thinking, he should really look at The Little Typer and Pie. Then that was Section V! (Really, many of the books in that series do this. The Reasoned Schemer for logic programming, The Little Prover for ACL2-style f
35.
by alew1 8y ago
We need a digital representation of real numbers. One possible choice is: a real number r is represented as the source code of a function which takes a natural number n to the nth digit of r. This seems no more like “cheating” (and to me, i
36.
by alew1 8y ago
I taught high school freshmen using HtDP for three years and loved it. Definitely works for high schoolers!
37.
by alew1 8y ago
Or maybe more analogous to “differentiable programming” or “logic programming”: a style of programming that supports a broader range of operations than “running.” Also, it’s not quite that you are using a DSL to encode Bayesian models tradi
38.
by alew1 8y ago
Great connection. Here's one way of looking at it: when you add stochasticity to your program, you make it easier to search the space of possible solutions (and you also relax the criteria for what it means to be "a solution"
39.
by alew1 8y ago
"Updating priors" is quite complex, and arguably describes every interesting task you'd want to do as a learner, reasoner, analyst, etc. One view of probabilistic programming languages is that they allow you to express your p
40.
by alew1 8y ago
It is, but the research that happens there is still funded in large part by public grants.
41.
by alew1 8y ago
Did the article change? The rest of the “hard to analyze” sentence you quoted above is “linear systems respond proportionally to changes in variables, whereas nonlinear systems have more complicated relationships.“
42.
by alew1 8y ago
I’ve seen 30 since signing up at the end of November. Not sure how typical that is, but my friends with MoviePass also see a bunch!
43.
by alew1 8y ago
I highly doubt they spent weeks training on the toy example. More like five minutes, probably. Again, that the weights learned (quickly) for the toy example can also be set by hand is not surprising and is evidence of a good toy example. Th
44.
by alew1 8y ago
This doesn’t seem like a particularly fair criticism. 1. As others have pointed out, the ImageNet experiment is presented as evidence that (as you’d expect) adding coordinate channels doesn’t affect performance on image classification tasks
45.
by alew1 8y ago
Question: in the list comprehension example, is `x` defined outside the loop comprehension? Will it overwrite an existing `x`? Or is it like `z` and only visible inside the loop comprehension? The latter would make more sense, I think.
46.
by alew1 8y ago
I’ve spent the last three years (straight out of college) teaching high school computer science. I’m headed to graduate school in the fall, but I’ve loved teaching, and think it’s a shame that fewer CS people go into the field. (Pay isn’t a
47.
by alew1 8y ago
After playing with this, I can make my brain hear either — and even in rhythmic patterns (Yanny, Laurel, Yanny, Laurel, Yanny, Yanny, Yanny, Laurel). It’s a neat reminder of how much “preprocessing” our brains do before we experience a soun
48.
by alew1 8y ago
Good response by aphyr. If I’m reading right, the original poster left because LLVM partnered with an organization designed to create internship opportunities for groups traditionally underrepresented in tech. In particular, LLVM agreed to
49.
by alew1 8y ago
I think here they just meant, show that no rational number is the square root of 2. Not rational = not expressible as a ratio. You assume that sqrt(2) = p/q for some integers p and q, then derive a contradiction.
50.
by alew1 8y ago
It seems likely that the two years of effort were essential in getting him to the point where he could have the breakthrough.
51.
by alew1 9y ago
An analogy that is used in the YouTube debate linked from the article is that of a scorekeeper and players of a game. The government is the scorekeeper, and operates according to rules we collectively agree on. Those rules can incentivize c
52.
by alew1 9y ago
Bootstrap is working on a high school physics curriculum for ninth graders, which integrates functional programming as a fundamental component. They only have a landing page up at the moment [0], but I attended a very cool workshop they gav
53.
by alew1 9y ago
Makes sense! I tend to take the "wrapping your point in uncertainty" approach, without all the ego-stroking. Nothing against ego-stroking; it just seems hard to make it feel genuine enough for it to work. Most comments (mine inclu
54.
by alew1 9y ago
I wonder if it could be done in a less grating way? I agree with the general idea of affirming the good in others' viewpoints, but I do find GP's phrasing patronizing. Why not something like: I agree, and I think there'd be a
55.
by alew1 9y ago
Not sure what his source is, but from the article: Andrew Clement, a computer scientist and expert in surveillance studies, has been mapping the NSA’s warrantless wiretapping activities since before Snowden’s disclosures. He strongly beli
56.
by alew1 9y ago
Hi all! Coded this up over the holidays, and curious to hear thoughts & suggestions! In order to find the "next word" w to say, the program uses a normalized set of word vectors and tries to maximize the product of the cosine
57.
Show HN: Can you think like a word vector? A game for exploring word embeddings(robotmindmeld.com)
8 points by alew1 9y ago|2 comments
58.
Show HN: Explore deep word embeddings in a fun, browser-based word game(robotmindmeld.com)
2 points by alew1 9y ago|0 comments
59.
by alew1 9y ago
Whether or not the economics are right, this author doesn’t understand the technology: > To explain: essentially, “blocks” are segregated, vast bundles of data in permanent communication with each other so that each block knows what the
60.
by alew1 9y ago
Thanks. I can see that this is a useful experiment, and I'm glad that the authors did it. It's just not clear to me that their model is new or needs a special name; isn't this single-stream approach just the simplest interpre

More ›