mxwsn

8 ms·

91.
by mxwsn 4y ago
Their extraction: (1) assumes the attacker knows the caption for some training images, and (2) primarily works on images duplicated 100x-3000x in the training dataset. Their attack does not succeed for any singleton images. Deduplicating ca
92.
by mxwsn 4y ago
That was posted on April fools. That's not the only one he's done, either.
93.
by mxwsn 4y ago
This can happen all the time in data science. Just add a few layers of bureaucracy and department separations between data gathering, cleaning/filtering, and analysis and it's easy to make the wrong conclusions on observed correla
94.
by mxwsn 4y ago
For all the talk of how doing a PhD is terrible, friendships in a college-like environment after undergrad was a bright point for me and many others. I moved and started a 9 to 5 last year and it's quite a change in lifestyle, and at 3
95.
by mxwsn 4y ago
Metamask and Brave wallets are free, popular and very easy to set up. They're considered hot wallets and not as secure as a hardware cold wallet like Ledger, but you'd have to wait for shipping and it isn't free. On screwing
96.
by mxwsn 4y ago
Perhaps before 1983, but there's a strong case that since the Orphan Drug Act was passed in the US, orphan drugs are overincentivized and a massive misallocation of resources [0]. Orphan drugs represent nearly half of FDA approved drug
97.
by mxwsn 4y ago
The author undersells himself - AlphaGo basically searches the game tree the same way he describes of GOFAI. Monte Carlo Tree Search is old yet essential. The neural network mainly improves the game evaluation heuristic, using function appr
98.
by mxwsn 4y ago
Can I ask how you use scientific language models professionally? Or do you have any articles/reviews on how they are being used, and how people see their potential and shortcomings?
99.
by mxwsn 4y ago
I love mate, but just one extra toilet trip is a substantial underestimate for many..
100.
by mxwsn 4y ago
In what sense is this a social network? From examples it looks like personal pages made in notion. But no friends list or commenting on other people's content?
101.
The Greatest Threat to AGI Is Narrow AI(argmax.ghost.io)
3 points by mxwsn 4y ago|0 comments
102.
by mxwsn 4y ago
The Nobel committee usually prefers to wait and evaluate longer-term impact, so I'd be quite surprised. CRISPR was obviously revolutionary in 2013 (imo, more than alphafold), and won the Nobel in 2020.
103.
by mxwsn 4y ago
From listening to a podcast or two, my impression is that the Public Bathroom Problem is an economic one. There were many paid public bathrooms in the past in the US, but they were deemed sexist and banned. Then cities didn't want to f
104.
by mxwsn 4y ago
I loosely mean how drosophila in biology is a relatively simple yet representative model that is easy to study that yields some insights about other systems.
105.
by mxwsn 4y ago
I wasn't familiar with this area, and it's so cool to hear about Moore's law progress on solving SAT, and how SAT is like the drosophila model. To me "automated reasoning" brings to mind reasoning with large languag
106.
by mxwsn 4y ago
The social default is in some fields where work is hard to reproduce is they still get into good journals and are believed up front. Usually this is fine. A lot of work that is hard to reproduce, but are relatively important or open up many
107.
by mxwsn 4y ago
I'm curious about your journey. In another thread you mentioned to me that you're currently debating leaving a faculty position for industry research. To go back and forth twice seems really unusual and I'm curious to hear mo
108.
by mxwsn 4y ago
"One of the ML mistakes that you will hurt you the most: not collecting the right features. If the information you need isn't in your data, you won't recover it with a better model." The word 'information' is d
109.
by mxwsn 4y ago
I work on ML+bio/chem at Genentech. I interviewed someone who deferred their asst. professorship and I think we're likely to hire them.
110.
by mxwsn 4y ago
I was in your shoes (undergrad class of 2015). My PhD journey was way more isolating, cynicism-inducing, political, yet also more successful than I could have imagined. I now have a nice researcher job in industry that has lots of agency, i
111.
by mxwsn 4y ago
A tip to ease anyone interested in making pixel art: instead of starting from scratch (very hard), for beginners it's much easier to take a high res image and down sampling it as a template. The down sampling algorithm will create many
112.
by mxwsn 4y ago
Variance is a critical lifeblood for card games, as demonstrated by the commander companion mechanic debacle, as a key lesson that mtg r&d has known but occasionally forgets. But not only at casual levels but many levels, mtg can defini
113.
by mxwsn 4y ago
Eth will hit a stable point of total supply, with both inflationary and deflationary mechanisms mentioned in other comments finding an equilibrium according to market forces. It won't deflate forever, so there won't be a centraliz
114.
by mxwsn 4y ago
I hate to be that person, but after reading all 7 posts I am starting to question if this is fake. I saw a family member go from diagnosis to death from stage 4 cancer and several things just feel off to me. The timeline of the posts, with
115.
by mxwsn 5y ago
Very amusing. Though, she concludes there's no difference based on average sleep quality, but the variance is sizably larger, and I personally care quite a bit about sleep quality consistency.
116.
by mxwsn 5y ago
You can also move your butt behind, over the rear wheel, to help stay grounded against the force of a strong press of the front brake.
117.
by mxwsn 5y ago
Dutch crunch is a very popular bread for sandwiches in the SF Bay area.
118.
by mxwsn 5y ago
I wonder if one can place mathematical concepts on a spectrum from discovery to invention? To me, the pythagorean theorem feels much more like a discovery of a "hidden" eternal truth that was once beyond our grasp. But to me, comp
119.
by mxwsn 5y ago
Oh wow, that's incredible. I didn't zoom in enough when I first looked, evidently!
120.
by mxwsn 5y ago
Nice! What area are you aiming for a PhD in?

More ›