breandan @breandan

student github.com/breandan/ [email protected] Joined September 2007

Tweets

4K
Followers

2K
Following

533
Likes

3K

Formal Languages and Automata Theory Papers @FormalLanguages

a month ago

Hamming distance between finite transducers Luc Dartois, Pierre-Cyrille Héam, Ismaël Jecker, Silvio Vescovo arxiv.org/abs/2604.25398 [𝚌𝚜.𝙵𝙻]

0 1 0 116 0

View Details

Formal Languages and Automata Theory Papers @FormalLanguages

8 months ago

Bringing Algebraic Hierarchical Decompositions to Concatenative Functional Languages. arxiv.org/abs/2510.12481

0 1 1 140 0

View Details

breandan @breandan

4 days ago

@credenzaclear2 Part of an onslaught of propaganda starting from early childhood to implant panpsychist and pantheistic beliefs into the collective subconscious. Large language models just add fuel to the dry tinder of neopaganism that has accumulated for decades unless properly contextualized.

breandan @breandan

5 months ago

The pervasive use of anthropomorphic characters in children's programs is part of an insidious effort to inculcate false compassion for nonhuman life, depersonalize the authentic human attributes it impersonates, and revive animistic tendencies dormant in the human consciousness.

0 0 0 330 1

0 0 0 63 0

View Details

Formal Languages and Automata Theory Papers @FormalLanguages

5 days ago

An Algebraic View of the Expressivity of Recurrent Language Models Franz Nowak, Ryan Cotterell, Reda Boumasmoud arxiv.org/abs/2606.01765 [𝚌𝚜.𝙵𝙻 𝚌𝚜.𝙲𝙻 𝚌𝚜.𝙻𝙶]

0 1 2 258 2

View Details

breandan @breandan

a week ago

@basedjensen Ascribing a conscience to machines would be a moral hazard, shifting the burden of responsibility to a nebulous other. This manufactures plausible deniability, shielding users and operators when taking unnecessary risks ("Just blame the AI!") and further diffusing responsibility.

0 0 0 49 0

View Details

breandan @breandan

a week ago

@basedjensen As for moral responsibility, this question remains squarely under the framework of double effect. As moral agents, we must not abdicate judgement for delegated actions, and likewise take reasonable precautions to avoid foreseeable harm when extending that agency through machines.

1 0 1 229 0

View Details

breandan @breandan

2 weeks ago

@_CLancellotti Arithmetic can be remarkably subtle! irma.math.unistra.fr/~schappa/NSch/… Even counting is quite difficult in practice. cs.toronto.edu/~meel/Papers/h…

0 0 1 244 0

View Details

Pope Leo XIV @Pontifex

2 weeks ago

Humanity, created by God in all its grandeur, is today facing a pivotal choice: either to construct a new Tower of Babel or to build the city in which God and humanity dwell together. In Jesus Christ, this humanity in its grandeur becomes the Way, the Truth and the Life, opening the path for each of us to grow toward fullness. #MagnificaHumanitas vatican.va/content/leo-xi…

1K 28K 178K 22.2M 38K

View Details

breandan @breandan

2 weeks ago

@jawwwn_ @PalmerLuckey The remarkable thing is, gender diversity in those engineering programs is often higher than their Western counterparts despite being just as, if not more competitive. I cannot stress how many well-adjusted women I have met from these places who never heard of affirmative action.

1 0 0 99 0

View Details

breandan @breandan

2 weeks ago

@jawwwn_ @PalmerLuckey I have long suspected this is part of a deliberate effort to erode STEM training. Those promoting soft skills in the West are fiercely pursuing technical research agendas back home. Visit any of the top universities in BRICS - you will not find DEI there. x.com/breandan/statu…

breandan @breandan

3 years ago

@loginn They must be doing something right. Our curriculum prefers teaching softer skills like engineering, which has value, but is much more tolerant of mediocrity. Dijkstra in particular, wrote some scathing remarks on the state of CS education in North America: cs.utexas.edu/users/EWD/tran…

0 0 3 3K 1

2 1 19 3K 2

View Details

breandan @breandan

2 weeks ago

@wavetossed @jawwwn_ @PalmerLuckey I have known some truly gifted engineers in the industry. There is another kind of brilliance you will find only in the corner of some forgotten department wrestling a decade-old problem with nothing but obsession to keep the fires burning. DEI mandates drove them mostly extinct.

0 0 0 81 0

View Details

breandan @breandan

2 weeks ago

Python users should experience considerably faster response times, thanks to a new Web Worker backend. Could be optimized even further with a WebGPU Python compiler. Next obstacle is to parallelize the CKY parser or implement a nonuniform intersection oracle. Suggestions welcome!

0 0 0 168 0

View Details

breandan @breandan

5 months ago

Updated the Python web demo to provide a more IDE-like look and feel. Currently, it just uses an n-gram reranker to fetch the top-k repairs. Our next goal is to port the transformer-based reranker to WebGPU, which yields a much more natural top-k ranking. tidyparse.github.io/python.html

1 0 1 465 1

View Details

breandan @breandan

2 years ago

Tidyparse is the seedling of a new structured editor that responds to every keystroke without any buttons or external configuration needed. To my knowledge, it is also the first sound and complete interactive proof assistant for the theory of bounded CFLs. tidyparse.github.io

3 1 27 5K 6

View Details

breandan @breandan

3 weeks ago

Remarkable how many crucial concepts needed for differentiable programming are anticipated by that '65 dissertation. She very clearly describes the BNF grammar, rewrite semantics, data structures, and computational primitives required to train most modern machine learning models.

0 2 1 198 1

View Details

breandan @breandan

3 weeks ago

@ch402 @ylecun @davidad @atilimgunes Interestingly, one of the first DSLs for automatic differentiation was developed by Sr. Mary Keller, who called it a "Program for Analytic Differentiation" (PAD) in her 1965 doctoral dissertation at @WisconsinCS (also one of the first PhDs awarded in CS): compcalc.github.io/public/keller/…

3 0 2 228 0

View Details

breandan @breandan

7 years ago

Interesting article on the origin of the term "differentiable programming". TLDR: popularized in 2018 by @ylecun, but seems to originate with @ch402's 2015 blog post, later abbreviated by @davidad. But I'm curious, what was wrong with "automatic differentiation"? cc: @atilimgunes

Shigekazu Ishihara @shigekzishihara

7 years ago

Where did "differentiable programming" come from? · Bonotake bonotake.github.io/deep%20learnin…

0 1 6 0 0

3 2 14 0 4

View Details

breandan @breandan

3 weeks ago

@ch402 @ylecun @davidad @atilimgunes @WisconsinCS In another related historical note, Bl. Francesco Faà di Bruno published in 1855 a generalization of the chain rule to higher order derivatives, with interesting applications to AD. I learned about this from @rxwei, who first taught me about it at the @diffprogramming workshop.

0 0 1 86 0

View Details

Formal Languages and Automata Theory Papers @FormalLanguages

a year ago

Uniform Membership for Hyperedge Replacement Grammars and Related Decision Problems. arxiv.org/abs/2501.08082

0 1 0 180 0

View Details

breandan @breandan

3 weeks ago

@ProgPapers @symbolica

0 0 0 34 0

View Details

breandan @breandan

3 weeks ago

Had a great time hanging out with the FLaNN folks in New Haven this week! Thanks to Andy, Bob, Dana, Lena, and all the great speakers for putting together a thought-provoking program. It was nice to hear about the cool research ideas people had during the brainstorming sessions.