darosati

6 ms·

1.
by darosati 2y ago
Hear hear
2.
by darosati 2y ago
I just want to push back on Academic Machine Learning is a (low quality with no novelty) paper mill and devaluing researchers efforts. To be clear ML research has “paper mill” problems but we should be careful that we don’t imply that there
3.
by darosati 2y ago
I would not say it’s impossible… my lab is working on this ( https://arxiv.org/abs/2405.14577 ) and though it’s far from mature - in theory some kind of resistance to downstream training isn’t impossible. I think under c
4.
Training-time domain authorization could be helpful for safety(lesswrong.com)
1 points by darosati 2y ago|0 comments
5.
Representation noising effectively prevents harmful fine-tuning on LLMs(arxiv.org)
1 points by darosati 2y ago|0 comments
6.
by darosati 4y ago
I don’t understand why very large neural networks can’t model causality in principal. I also don’t understand the argument that even if NNs can model causality in principal they are unlikely to do so in practice (things I’ve heard: spurious
7.
Ask HN: Are there any good Diff tools for Jupyter Notebooks?
51 points by darosati 4y ago|27 comments
8.
by darosati 4y ago
My research involve applying Popper's epistemology to natural language processing. So I am quite involved in this. As far as I can tell, almost all of what Popper tried to do with quantification measures of information are exactly what
9.
by darosati 5y ago
Ah ya that makes total sense. I will see what we can do to make that more obvious and linked to our paper reports thanks for the suggestion.
10.
by darosati 5y ago
Hey there, I work at scite and yes it defenitly can. In general we have "who is funding what research" data here ( https://scite.ai/institutions ). We don't really have a freindly way for you to produce a heatm
11.
by darosati 5y ago
Disclaimer: I work at scite. We, at scite, actually designed scite to help folks see beyond the "one big claim per paper" model. Specifically, we designed scite to present show how specific results were either supported or contras
12.
by darosati 5y ago
Thanks! I am glad you are finding it useful (I work at scite)
13.
Scite Reference Check: an easy way to check the reliability of your references(medium.com)
1 points by darosati 6y ago|1 comments
14.
Visualize Scientific Debates(scite.ai)
2 points by darosati 6y ago|0 comments
15.
Show HN: Visualize Supporting and Disputing Citations(scite.ai)
3 points by darosati 6y ago|0 comments
16.
by darosati 6y ago
https://scite.ai/reports/brain-drain-the-mere-presence-2NEep... Interesting support for this article
17.
Scite Orcid Integration(medium.com)
1 points by darosati 6y ago|0 comments
18.
Show HN: Scientific Support Badges(scite.ai)
3 points by darosati 6y ago|0 comments
19.
Scite Reference Check: an easy way to check the reliability of your references(medium.com)
2 points by darosati 6y ago|0 comments
20.
by darosati 6y ago
REMOTE(US) PAID INTERN: At scite.ai we are hiring someone to help us get traction! We are a company focused on making science more relabile through NLP, Biblometrics, more. https://scite.ai/jobs/marketing-intern
21.
Ask HN: Web app stack from scratch what technologies would you choose?
1 points by darosati 6y ago|2 comments
22.
by darosati 7y ago
I am one of the folks working on Valence, i'd love feedback on what we are doing. Valence attempts to solve for performance management by giving folks a declarative way of setting performance objectives and having a controller go and m
23.
Show HN: Valence, Declarative Performance for Kubernetes(valence.net)
17 points by darosati 7y ago|1 comments
24.
Show HN: PyTill: A Python Package for Till Mobile(blog.manifold.co)
9 points by darosati 8y ago|0 comments