@ysawej.bsky.social
Deep Learning, NLP, multimodal models and all things neural network. Also, math guy and computer scientist.
Reposted
A city of drivers expands, relentlessly and inevitably, into sprawl.
A city of pedestrians, cyclists, and transit-users settles, gradually and elegantly, into itself.
A city of pedestrians, cyclists, and transit-users settles, gradually and elegantly, into itself.
January 6, 2025 at 8:57 PM
A city of drivers expands, relentlessly and inevitably, into sprawl.
A city of pedestrians, cyclists, and transit-users settles, gradually and elegantly, into itself.
A city of pedestrians, cyclists, and transit-users settles, gradually and elegantly, into itself.
Reposted
Still the best two sentences on climate action.
December 29, 2024 at 5:42 PM
Still the best two sentences on climate action.
Reposted
An illustrated guide to never learning anything
December 25, 2024 at 12:26 AM
An illustrated guide to never learning anything
Reposted
“The best part of waking up, is your house being quiet and your kids not tearing shit up”
November 29, 2024 at 11:22 PM
“The best part of waking up, is your house being quiet and your kids not tearing shit up”
Reposted
The next coffee drinker whom awakens, understands the joy, of a hot pot ready and waiting.
November 29, 2024 at 11:34 AM
The next coffee drinker whom awakens, understands the joy, of a hot pot ready and waiting.
Reposted
There is no greater joy than everyone in the house being asleep as you drink your morning coffee.
November 29, 2024 at 11:00 AM
There is no greater joy than everyone in the house being asleep as you drink your morning coffee.
Reposted
You’re still arguing about tabs vs. spaces? May I present…
December 25, 2024 at 6:37 PM
You’re still arguing about tabs vs. spaces? May I present…
Reposted
I didn't embark in machine learning thinking of it as an ideological project to disenfranchise human beings.
But we need to face reality, machine learning can easily become the driver of such change, shifting power structures.
We, actors of tech, can modulate this effect.
1/3
But we need to face reality, machine learning can easily become the driver of such change, shifting power structures.
We, actors of tech, can modulate this effect.
1/3
December 9, 2024 at 11:58 PM
I didn't embark in machine learning thinking of it as an ideological project to disenfranchise human beings.
But we need to face reality, machine learning can easily become the driver of such change, shifting power structures.
We, actors of tech, can modulate this effect.
1/3
But we need to face reality, machine learning can easily become the driver of such change, shifting power structures.
We, actors of tech, can modulate this effect.
1/3
📌
December 21, 2024 at 1:42 AM
📌
Reposted
"Public investment" vs "Wasteful Subsidy." The only problem with this clever Singer cartoon is that some people might actually not get that it’s sarcastically illustrating the perception problem, NOT telling the truth. Just in case it actually needs to be said, the truth is the opposite.
December 18, 2024 at 8:21 AM
"Public investment" vs "Wasteful Subsidy." The only problem with this clever Singer cartoon is that some people might actually not get that it’s sarcastically illustrating the perception problem, NOT telling the truth. Just in case it actually needs to be said, the truth is the opposite.
Reposted
Dynamic programming alternatives to dynamic programming for optimal control. Replace your Bellman equation with backpropagation. www.argmin.net/p/twirling-t...
Twirling Towards Freedom
Avoiding dynamic programming with dynamic programming
www.argmin.net
November 21, 2024 at 3:37 PM
Dynamic programming alternatives to dynamic programming for optimal control. Replace your Bellman equation with backpropagation. www.argmin.net/p/twirling-t...
Bookmark
Meta's Apollo Multimodal Models
- 1.5B, 3B and 7B model checkpoints
- Can comprehend up-to 1 hour of video
- Temporal reasoning & complex video question-answering
- Multi-turn conversations grounded in video content
Project: apollo-lmms.github.io
Models: huggingface.co/Apollo-LMMs
- 1.5B, 3B and 7B model checkpoints
- Can comprehend up-to 1 hour of video
- Temporal reasoning & complex video question-answering
- Multi-turn conversations grounded in video content
Project: apollo-lmms.github.io
Models: huggingface.co/Apollo-LMMs
December 16, 2024 at 5:08 PM
Bookmark
Bookmark
Hellinger and Wasserstein are the two main geodesic distances on probability distributions. While both minimize the same energy, they differ in their interpolation methods: Hellinger focuses on density, whereas Wasserstein emphasizes position displacements.
December 16, 2024 at 6:33 AM
Bookmark
Bookmark
This paper by Hornik et al demonstrates the *uniform* approximation universality of 2-layer MLPs with sigmoid activation functions, leveraging that sinusoids can approximate any function through Fourier expansion.
www.cs.cmu.edu/~epxing/Clas...
www.cs.cmu.edu/~epxing/Clas...
December 16, 2024 at 6:31 AM
Bookmark
Reposted
boundary2 just published a forum on the gordian knot of finance, where @stefeich.bsky.social, @aminsamman.bsky.social, @thisblue.bsky.social, Janet Roitman, Dick Bryan, and myself reflect on the infuriating hold of finance on economic policy (and how to break it)
www.boundary2.org/the-gordian-...
www.boundary2.org/the-gordian-...
the gordian knot of finance | special issue
This special issue is hosted by the "Finance and Fiction" dossier of the b2o Review, which is edited by Arne De Boever and Mikkel Krause Frantzen. Volume 6, Issue 1 (December 2024) Special Issue: Th...
www.boundary2.org
December 12, 2024 at 10:49 PM
boundary2 just published a forum on the gordian knot of finance, where @stefeich.bsky.social, @aminsamman.bsky.social, @thisblue.bsky.social, Janet Roitman, Dick Bryan, and myself reflect on the infuriating hold of finance on economic policy (and how to break it)
www.boundary2.org/the-gordian-...
www.boundary2.org/the-gordian-...
Reposted
The tech bro fascination with eugenics is so on brand. The idea that you could make accurate, actionable predictions from individual genotypes (with all their complex, non-linear interactions) from averaged, linearly modeled population-level genomic statistics is just another big data fantasy
December 12, 2024 at 4:13 PM
The tech bro fascination with eugenics is so on brand. The idea that you could make accurate, actionable predictions from individual genotypes (with all their complex, non-linear interactions) from averaged, linearly modeled population-level genomic statistics is just another big data fantasy
Reposted
Tech bros know what every parent wants: children who sleep less
December 11, 2024 at 2:39 PM
Tech bros know what every parent wants: children who sleep less
Reposted
noice!
December 9, 2024 at 10:49 AM
noice!
Reposted
Insane lack of transparency from OpenAI. Saying the models people use are different than the "final" evals they released. Do better. Especially enterprises will move elsewhere because of this stuff.
December 6, 2024 at 10:00 PM
Insane lack of transparency from OpenAI. Saying the models people use are different than the "final" evals they released. Do better. Especially enterprises will move elsewhere because of this stuff.
Reposted
Zuck is developing 2GW+ data center.
"Last big AI update of the year:
• Meta AI now has nearly 600M monthly actives
• Releasing Llama 3.3 70B text model that performs similarly to our 405B
• Building 2GW+ data center to train future Llama models
Next stop: Llama 4. Let's go!"
"Last big AI update of the year:
• Meta AI now has nearly 600M monthly actives
• Releasing Llama 3.3 70B text model that performs similarly to our 405B
• Building 2GW+ data center to train future Llama models
Next stop: Llama 4. Let's go!"
December 6, 2024 at 6:31 PM
Zuck is developing 2GW+ data center.
"Last big AI update of the year:
• Meta AI now has nearly 600M monthly actives
• Releasing Llama 3.3 70B text model that performs similarly to our 405B
• Building 2GW+ data center to train future Llama models
Next stop: Llama 4. Let's go!"
"Last big AI update of the year:
• Meta AI now has nearly 600M monthly actives
• Releasing Llama 3.3 70B text model that performs similarly to our 405B
• Building 2GW+ data center to train future Llama models
Next stop: Llama 4. Let's go!"
Reposted
Learning about quantization suffixes while `ollama pull llama3.3` download completes (fyi, quantization for the default 70b is q4_K_M)
• make-ggml .py: github.com/ggerganov/ll...
• pull request: github.com/ggerganov/ll...
• make-ggml .py: github.com/ggerganov/ll...
• pull request: github.com/ggerganov/ll...
December 7, 2024 at 1:09 AM
Learning about quantization suffixes while `ollama pull llama3.3` download completes (fyi, quantization for the default 70b is q4_K_M)
• make-ggml .py: github.com/ggerganov/ll...
• pull request: github.com/ggerganov/ll...
• make-ggml .py: github.com/ggerganov/ll...
• pull request: github.com/ggerganov/ll...
Bookmark
Thank you for following!
I'm a first-year doctoral student at the University of Tokyo.
I have published several papers related to speech synthesis and source separation, and my current research focuses on target speaker extraction.
Feel free to contact me!
Nice to meet you!
I'm a first-year doctoral student at the University of Tokyo.
I have published several papers related to speech synthesis and source separation, and my current research focuses on target speaker extraction.
Feel free to contact me!
Nice to meet you!
December 6, 2024 at 2:49 AM
Bookmark
Reposted
Will be at #NeurIPS2024 Dec 10-13. Looking forward to run into everyone in #AI all at once🤞 The 2019 Vancouver one was the largest conf I ever attended--not sure how they plan to cram even more this time.. 😱
[Will be at our poster on 12/11 morning openreview.net/forum?id=kPB... ]
[Will be at our poster on 12/11 morning openreview.net/forum?id=kPB... ]
Chain of Thoughtlessness? An Analysis of CoT in Planning
Large language model (LLM) performance on reasoning problems typically does not generalize out of distribution. Previous work has claimed that this can be mitigated with chain of thought...
openreview.net
November 30, 2024 at 2:35 PM
Will be at #NeurIPS2024 Dec 10-13. Looking forward to run into everyone in #AI all at once🤞 The 2019 Vancouver one was the largest conf I ever attended--not sure how they plan to cram even more this time.. 😱
[Will be at our poster on 12/11 morning openreview.net/forum?id=kPB... ]
[Will be at our poster on 12/11 morning openreview.net/forum?id=kPB... ]
Reposted
📣I'm hiring PhD interns for combined theory+empirical projects in: exploration in post-training, multi-task learning in autoregressive models, distillation, reasoning beyond CoT.
Apply on the link below. If you're at #NeurIPS2024, message me to chat.
jobs.careers.microsoft.com/global/en/jo...
Apply on the link below. If you're at #NeurIPS2024, message me to chat.
jobs.careers.microsoft.com/global/en/jo...
December 5, 2024 at 3:42 PM
📣I'm hiring PhD interns for combined theory+empirical projects in: exploration in post-training, multi-task learning in autoregressive models, distillation, reasoning beyond CoT.
Apply on the link below. If you're at #NeurIPS2024, message me to chat.
jobs.careers.microsoft.com/global/en/jo...
Apply on the link below. If you're at #NeurIPS2024, message me to chat.
jobs.careers.microsoft.com/global/en/jo...