dimatura

12 ms·

121.
by dimatura 3y ago
So I see that the name is derived from an esperanto word (apparently meaning "flatter") but in spanish it would also mean a small burp ("flato" being an informal term for burp). Which I still like as a name ;)
122.
by dimatura 3y ago
Haha, I'm one of the authors, although most of the credit goes to the other author (David Fouhey). Despite being in the form of a joke, this sheet is actually functional, for anyone wondering. Like the one in this post, it's infer
123.
by dimatura 3y ago
In the late 2000s I worked at a lab that used subversion, which I also found annoying after getting used to git. My hack was to use git-svn - it let you locally use git, and eventually sync your work back to svn. Worked like a charm. The re
124.
by dimatura 3y ago
Yes, back in the day I saw people doing this with Maple, glad there's open source options now. There's a cool library (also using Sympy) to do this kind of thing in robotics and computer vision applications, symforce https:/
125.
by dimatura 3y ago
Yeah, definitely opaque. If I had to guess it sort of sounds like a code optimization that resulted in a numerical error, but only in some GPUs or CUDA versions. I've seen that sort of issue happen a few times in the pytorch framework,
126.
by dimatura 3y ago
As someone in CV/ML with a background in cognitive science, and partial to the work of Lakoff and others in regards to the role of embodiment, it does irk me how loose papers in vision can be with the term. In this case specifically, i
127.
by dimatura 3y ago
Right, that's why the "noisy TV problem" is a "problem" - it makes the simple model of curiosity used in the research less effective, because it causes agents to get stuck watching TV instead of exploring and advanc
128.
by dimatura 3y ago
In robotics, it's pretty widely used to refer to a 6-degree of freedom pose or relative pose in space, for example it's widely used in ROS, a de facto standard https://docs.ros.org/en/melodic/api/geo
129.
by dimatura 3y ago
I like this summary of Lie theory in the context of robotics: https://arxiv.org/pdf/1812.01537.pdf
130.
by dimatura 3y ago
Yeah, I'm happy there's a lot of development in this area - even if it's fueled by the LLM frenzy, good nearest neighbor search solutions are useful in a lot of domains. Though I worked a little bit on this problem over 10 ye
131.
by dimatura 3y ago
Yeah, I'm middle-aged and ride around on an electric skateboard, including a daily but very short commute. I go out of my way (literally) to not inconvenience either pedestrians or drivers, so have had only a couple of drivers yell at
132.
by dimatura 3y ago
2D markers that use color to encode extra bits of information do exist actually, see for example http://josephdegol.com/docs/ChromaTag_ICCV17_Paper.pdf . So it's not a bad idea. But as other commenters mentioned, f
133.
by dimatura 3y ago
A bit of a tangent, but reminds me of this article, RenderGAN [1]. The main point of the article was about using 3D rendering + GAN to generate convincing synthetic data, and their application was creating training data to detect "bar
134.
by dimatura 3y ago
I don't know if it will be significantly different than what Google Translate does, but I would give the major cloud vendors (Google, Amazon, Microsoft and I guess OpenAI/ChatGPT) OCR services a shot. It's pretty simple and c
135.
by dimatura 3y ago
awesome, how did I not find this :D
136.
by dimatura 3y ago
Thanks, I'll check it out.
137.
by dimatura 3y ago
I already use jupytext to store notebooks as code but the improved state management and notebook-as-app features are pretty compelling and I'm trying it out. Unfortunately, I'm quite used to very specific vim keybindings in Jupyte
138.
by dimatura 3y ago
Any recommendations for an embedded embedding database (heh)? Embedded, as in sqlite. For smaller-scale problems, but hopefully more convenient than say, LMDB + FAISS.
139.
by dimatura 3y ago
For those two in particular? You'd definitely get a better result with an ML model such as a convolutional neural net. In some sense, using an image similarity query is a kind of ML model - nearest neighbor - which can work in some sce
140.
by dimatura 3y ago
Automatic differentiation isn't just for machine learning, it's useful in several other applications where you have an optimization problem built of many variables interacting in differentiable ways. One possible guess could be st
141.
by dimatura 3y ago
This reminds me of something I noticed on this year's "Summer of Baja Blast" promotion - to make an entry, all you had to do was take a picture of a bottle with your phone - it wasn't scanning any barcodes or QR codes. I
142.
by dimatura 3y ago
awesome! I actually have this synth on ios as well. Can't have enough FM.
143.
by dimatura 3y ago
Almost certainly the legendary Casio VL-1 https://en.wikipedia.org/wiki/Casio_VL-1 . Featured in a quite a few famous and not-so-famous songs.
144.
by dimatura 3y ago
I encountered Close to the Machine by chance in my university's library, many years ago. I flipped through the first few pages and was quickly hooked. Definitely a great read.
145.
by dimatura 3y ago
Surprisingly, this display is miles better than the ones we keep seeing in new synths from some of the major manufacturers, like Roland and Korg.
146.
by dimatura 3y ago
I'm not sure about how the longevity of this business model. It seems to me that the quality of generative models will continuously improve over time (in part by incorporating similar kinds of "bad output filters"), making th
147.
by dimatura 3y ago
The Coral line of products targets embedded/IoT inference applications, and are on the low end as far as processing power goes. AFAICT the recent surge of demand, mostly fueled by LLMs, is for GPUs on the opposite end of processing pow
148.
by dimatura 3y ago
Well, that idea is one of the motivations behind the paper, which is itself a throwback to earlier ideas about the "language of thought", a hypothetical language (certainly not ordinary natural language, and probably more like a p
149.
by dimatura 3y ago
I disagree that this is an "unfounded hot take". It's far from a rare opinion on cognitive science, and if I had to guess it's probably the mainstream opinion (I can't really back that up with citations because I ha
150.
by dimatura 3y ago
This is really interesting. The title is referencing the "Language of Thought" hypothesis from early cognitive psychology, that posited thought consisted of symbol manipulation akin to computer programs. The same idea was behind w

More ›