tsumnia

9 ms·

121.
by tsumnia 2y ago
I talk briefly about this in my video on Heuristic Functions (I won't link just to stop shameless plugging), but its a combination of funding, knowledge, and team. Scheduling problems are a specific type of Optimization problem and man
122.
by tsumnia 2y ago
Shamelessly Plugging my own videos on Linear Programs and Meta-heuristic Algorithms to solve Linear Assignment Problems. The Linear Assignment Problem: https://www.youtube.com/watch?v=dMJpCnocmGk Simulated Annealing: https
123.
by tsumnia 2y ago
Neat, but I wish Meta would just say what this really is - "please give us some In the Wild data to further train our models on". I did the same technique years ago for estimating ages. Person uploads an image, helps align 10% of
124.
by tsumnia 2y ago
I will admit over reliance on AI is a major issue that we're coming to terms with right now. However to invoke playing devil's advocate, a person over relying on stimulants can also be a bad thing. In moderation, AI can be fine an
125.
by tsumnia 2y ago
I'm not against that approach (though I am a teacher so guilty as charged). Toy examples help teach a concept and it helps when the example is relevant to the learner's interest. However at some point, we can't design real wo
126.
by tsumnia 2y ago
Oh absolutely, I once saw research exploring how terrorist groups were using World of Warcraft emotes as their steganography and I show my students how to do simple least significant bit hiding when I teach them about image processing. Unle
127.
by tsumnia 2y ago
I can't speak to Caltech, but it sounds like what you're describing is a perfect example of a "worked example". The idea is that you are walking through an algorithm, piece by piece even to the most basic levels. Its the
128.
by tsumnia 2y ago
Obligatory XKCD: https://xkcd.com/538/
129.
by tsumnia 2y ago
They don't need to be obscene - I think a good example would be non-sparring martial arts like Aikido or Tai Chi. There isn't really an incentive to pay to watch top performers in the arts demonstrate technique beyond a handful of
130.
by tsumnia 2y ago
I sort of disagree, as we've seen a significant increase in small scope tools like the FlipperZero, Arduinos, and Raspberry Pis. Each one has a slew of shields/extensions/whatever that subculture calls their plug and play opt
131.
by tsumnia 2y ago
Check out a lot of DisneyResearchHub's videos on YouTube. They are regularly posting research on object/face tracking and projections. That said as you mentioned faces are pretty stereotyped. We "solved" face mapping two
132.
by tsumnia 2y ago
Believe me, I get it. BUT, and even I hate saying this, if both sides continue to resort to "the other side is lying" - aka people claiming 2020 was rigged and people claiming deportation/crime numbers are under-reported. I&#
133.
by tsumnia 2y ago
While I can recognize the sentiment, I do err on the side of caution moving into the post-COVID world. Once you start assuming "the other side could be lying" then its going to be pretty difficult for you to ever accept whatever t
134.
by tsumnia 2y ago
Well OBVIOUSLY except for Detritus! Things is too many things to count.
135.
by tsumnia 2y ago
And once Nobby says he likes Tramway, everyone realizes Vetinari was right all along XD (I'm just glad someone got the reference)
136.
by tsumnia 2y ago
"A thing should be a thing. It should not be a bunch of things pretending to be a single thing. With nodes you have to pretend that a collection of things is a single thing." Just want to say this line was great, very Terry Pratch
137.
by tsumnia 2y ago
I'm not familiar with Sung's videos, but after a quick perusal of his thumbnails I have some knowledge in the various cognitive science concepts he covers (Cognitive Load Theory, Flow, Mindmaps). I didn't really dig into educ
138.
by tsumnia 2y ago
I don't personally use Anki, not for any major reasons or ideological disagreements but mostly out of a "I do something else to learn" methodology. I've always liked the concept of Anki, flashcard learning, and needing s
139.
by tsumnia 2y ago
The year of the leopard print futon. I thought I was the coolest kid in the world because my bed could fold up into a couch. Funny enough, I still sleep great on futons. These days if I lie down on a futon, go ahead and assume I'm abou
140.
by tsumnia 2y ago
The 'plastic rock' formations off Brazil [1] may be something? Like if advanced civilizations incorporate some element of time/pressure/environments to something, the plastic that has melted, merged, and reformed into ro
141.
by tsumnia 2y ago
> And to echo your argument: in familiar environments those shortcuts are often a good idea! Only to continue to reaffirm the original post, this was some of the basis for my dissertation. Lower-level practice, or exposure to tons of int
142.
by tsumnia 2y ago
Firefox ran it, but very broken.
143.
by tsumnia 2y ago
Military food is a really fascinating logistics problem - how do you feed people, potentially across a long period while engaging in calorically demanding tasks? Foodstuffs need to last as long as possible because supply lines are targeted
144.
by tsumnia 2y ago
Despite picking up computers at a young age, that was in 1995, nearly 3 decades after this moment. It still blows my mind at what we've been able to do with the power to simulate and visualize everything around us. It's something
145.
by tsumnia 2y ago
Encarta 95 was one of the greatest pieces of software history created in my opinion. I know I would spend hours playing the adventure game in it, swapping between the map and articles to answer all the questions. There's been some atte
146.
by tsumnia 2y ago
All of those Choose Your Own Adventure books really built my interest in stories of all kind. I'd juggle 3-4 different by using my fingers as bookmarks until I couldn't hold the book any longer and had to pick ONE storyline.
147.
by tsumnia 2y ago
Tons, but there are notable standouts. - The Hitchhikers Guide to the Galaxy by Douglas Adam - Spent all of senior year in HS reading the entire series - Harrison Bergeron by Kurt Vonnegut - Maybe it was because I mentally compared Harrison
148.
by tsumnia 2y ago
The issue is that an LLM by itself does not try to be correct or incorrect, only respond with tokens that have a high probability of appearing next. Optimal only becomes a metric for consideration when humans were added to rate the "go
149.
by tsumnia 2y ago
Don't think of it as "less optimal", but rather other next tokens that have smaller, but still high, probabilties of being selected. If your largest next token has a probability of (arbitrarily) 25% of being selected, this do
150.
by tsumnia 2y ago
Nicely done and keep up the practice. I recall during my Masters needing to translate facial landmark points from a Cartesian coordinate system into points that could would appear on their respective images. It wasn't for anything majo

More ›