Sadamori Kojaku
@skojaku.bsky.social
Assist. Prof. in 🇺🇸 . Interested in Network Science and Computer Science. Love Sailing, Kendo, Tennis, and Cooking Chinese and Italian. My favorite book: 11ぴきのネコ.
GPT Inference (greedy, top-k, beam search, temperature control).
November 29, 2025 at 11:27 AM
GPT Inference (greedy, top-k, beam search, temperature control).
word2vec: from the structural linguistic perspective. Written by me. Illustrated by Nano Banana Pro.
November 27, 2025 at 8:51 PM
word2vec: from the structural linguistic perspective. Written by me. Illustrated by Nano Banana Pro.
Reposted by Sadamori Kojaku
🧵1/
A bit late to the party, but I want to introduce a pet project of mine—and my first single-author paper! 🎉
It tackles something that has quietly frustrated me for years: how we measure the similarity of two clusterings/community structures.
A bit late to the party, but I want to introduce a pet project of mine—and my first single-author paper! 🎉
It tackles something that has quietly frustrated me for years: how we measure the similarity of two clusterings/community structures.
November 24, 2025 at 8:58 PM
🧵1/
A bit late to the party, but I want to introduce a pet project of mine—and my first single-author paper! 🎉
It tackles something that has quietly frustrated me for years: how we measure the similarity of two clusterings/community structures.
A bit late to the party, but I want to introduce a pet project of mine—and my first single-author paper! 🎉
It tackles something that has quietly frustrated me for years: how we measure the similarity of two clusterings/community structures.
Nano Banana Pro unbelievably lowers the entry barrier for people interested in creating short Manga. This blog post describes the recipe (in Japanese). I get this Manga version of my lecture note on the position encoding in transformers, **in the first shot** 😮.
note.com/konho/n/na8d...
note.com/konho/n/na8d...
November 24, 2025 at 6:09 PM
Nano Banana Pro unbelievably lowers the entry barrier for people interested in creating short Manga. This blog post describes the recipe (in Japanese). I get this Manga version of my lecture note on the position encoding in transformers, **in the first shot** 😮.
note.com/konho/n/na8d...
note.com/konho/n/na8d...
海南鶏飯を一日かけて作ったのだけど、子供は食べないばかりかぐちゃぐちゃにされて心が折れた。１０分でできるパスタ出してあげるよ。次から🫠。
November 24, 2025 at 11:03 AM
海南鶏飯を一日かけて作ったのだけど、子供は食べないばかりかぐちゃぐちゃにされて心が折れた。１０分でできるパスタ出してあげるよ。次から🫠。
Reposted by Sadamori Kojaku
I'm looking for a PhD student to join my lab at UAlbany for Fall 2026!
We are data scientists studying the creation, flow, and impact of information. How do scientists collaborate? How does AI shape information access? What drives local innovation?
Want to answer these questions? details below!
We are data scientists studying the creation, flow, and impact of information. How do scientists collaborate? How does AI shape information access? What drives local innovation?
Want to answer these questions? details below!
November 5, 2025 at 6:35 PM
I'm looking for a PhD student to join my lab at UAlbany for Fall 2026!
We are data scientists studying the creation, flow, and impact of information. How do scientists collaborate? How does AI shape information access? What drives local innovation?
Want to answer these questions? details below!
We are data scientists studying the creation, flow, and impact of information. How do scientists collaborate? How does AI shape information access? What drives local innovation?
Want to answer these questions? details below!
Reposted by Sadamori Kojaku
🚨 New paper on ArXiv:
“Uncertainty quantification and posterior sampling for network reconstruction”
TL;DR; We present an efficient method to sample the entire ensemble of possible network reconstructions that are compatible with an indirect observation, e.g. a dynamics.
Short thread: 1/N
“Uncertainty quantification and posterior sampling for network reconstruction”
TL;DR; We present an efficient method to sample the entire ensemble of possible network reconstructions that are compatible with an indirect observation, e.g. a dynamics.
Short thread: 1/N
March 12, 2025 at 11:57 AM
🚨 New paper on ArXiv:
“Uncertainty quantification and posterior sampling for network reconstruction”
TL;DR; We present an efficient method to sample the entire ensemble of possible network reconstructions that are compatible with an indirect observation, e.g. a dynamics.
Short thread: 1/N
“Uncertainty quantification and posterior sampling for network reconstruction”
TL;DR; We present an efficient method to sample the entire ensemble of possible network reconstructions that are compatible with an indirect observation, e.g. a dynamics.
Short thread: 1/N
TLDR: communities are undetectable even if they are because of the variabilities of network topology. We look further into edge weights, showing that a higher variability in edge weights makes communities harder to be detected. Our results inform when to use edge weights and better edge reweighing.
When are communities detectable in weighted networks? You can find the answer in our latest preprint. Great collaboration with @filipisilva.bsky.social @skojaku.bsky.social arxiv.org/abs/2511.00214
Detectability threshold in weighted modular networks
We study the necessary condition to detect, by means of spectral modularity optimization, the ground-truth partition in networks generated according to the weighted planted-partition model with two eq...
arxiv.org
November 4, 2025 at 6:30 PM
TLDR: communities are undetectable even if they are because of the variabilities of network topology. We look further into edge weights, showing that a higher variability in edge weights makes communities harder to be detected. Our results inform when to use edge weights and better edge reweighing.
Reposted by Sadamori Kojaku
Is #innovation born or made? Why do some places keep innovating while others fall behind? Is innovation the product of cities—or of deeper laws?
Our new paper in npj Complexity explores these questions through the lens of a century of US patent data.🧵👇
nature.com/articles/s44...
Our new paper in npj Complexity explores these questions through the lens of a century of US patent data.🧵👇
nature.com/articles/s44...
November 3, 2025 at 5:05 PM
Is #innovation born or made? Why do some places keep innovating while others fall behind? Is innovation the product of cities—or of deeper laws?
Our new paper in npj Complexity explores these questions through the lens of a century of US patent data.🧵👇
nature.com/articles/s44...
Our new paper in npj Complexity explores these questions through the lens of a century of US patent data.🧵👇
nature.com/articles/s44...
Reposted by Sadamori Kojaku
When are communities detectable in weighted networks? You can find the answer in our latest preprint. Great collaboration with @filipisilva.bsky.social @skojaku.bsky.social arxiv.org/abs/2511.00214
Detectability threshold in weighted modular networks
We study the necessary condition to detect, by means of spectral modularity optimization, the ground-truth partition in networks generated according to the weighted planted-partition model with two eq...
arxiv.org
November 4, 2025 at 3:52 PM
When are communities detectable in weighted networks? You can find the answer in our latest preprint. Great collaboration with @filipisilva.bsky.social @skojaku.bsky.social arxiv.org/abs/2511.00214
Wanted to create an interactive *public* visualization for networks/embeddings?
I created a minimum GitHub Page template for HeliosWeb visualization. @filipisilva.bsky.social
👉 Demo: shorturl.at/m9CnR
👉 Repo: shorturl.at/HJ2gn
👉 HeliosWeb: shorturl.at/H9bJU
#dataviz #datascience #python
I created a minimum GitHub Page template for HeliosWeb visualization. @filipisilva.bsky.social
👉 Demo: shorturl.at/m9CnR
👉 Repo: shorturl.at/HJ2gn
👉 HeliosWeb: shorturl.at/H9bJU
#dataviz #datascience #python
GitHub - skojaku/myheliosweb-template
Contribute to skojaku/myheliosweb-template development by creating an account on GitHub.
github.com
November 3, 2025 at 10:19 AM
Wanted to create an interactive *public* visualization for networks/embeddings?
I created a minimum GitHub Page template for HeliosWeb visualization. @filipisilva.bsky.social
👉 Demo: shorturl.at/m9CnR
👉 Repo: shorturl.at/HJ2gn
👉 HeliosWeb: shorturl.at/H9bJU
#dataviz #datascience #python
I created a minimum GitHub Page template for HeliosWeb visualization. @filipisilva.bsky.social
👉 Demo: shorturl.at/m9CnR
👉 Repo: shorturl.at/HJ2gn
👉 HeliosWeb: shorturl.at/H9bJU
#dataviz #datascience #python
Reposted by Sadamori Kojaku
A semi-regular reminder: My colleagues and I are running an interdisciplinary online seminar series called Complex Systems Research Exchange (CREx🦖).
sites.google.com/view/cxrex
sites.google.com/view/cxrex
November 3, 2025 at 10:13 AM
A semi-regular reminder: My colleagues and I are running an interdisciplinary online seminar series called Complex Systems Research Exchange (CREx🦖).
sites.google.com/view/cxrex
sites.google.com/view/cxrex
Snakemake is essential for reproducible research—it organizes code, and acts as my second brain about code and data relations.
A downside is its maintenance cost. It can be easily convoluted as a project progresses. This tool might resolve that.
arxiv.org/html/2505.02...
A downside is its maintenance cost. It can be easily convoluted as a project progresses. This tool might resolve that.
arxiv.org/html/2505.02...
Snakemaker: Seamlessly transforming ad-hoc analyses into sustainable Snakemake workflows with generative AI
arxiv.org
November 2, 2025 at 10:02 PM
Snakemake is essential for reproducible research—it organizes code, and acts as my second brain about code and data relations.
A downside is its maintenance cost. It can be easily convoluted as a project progresses. This tool might resolve that.
arxiv.org/html/2505.02...
A downside is its maintenance cost. It can be easily convoluted as a project progresses. This tool might resolve that.
arxiv.org/html/2505.02...
I've been playing with Ollama Cloud that has a free cloud Qwen 3 Vision Language model, which is surprisingly good. I put together a small app that automatically renames screenshot files based on their content, which has been super helpful for organizing them.
github.com/skojaku/scre...
github.com/skojaku/scre...
GitHub - skojaku/screenshot-renamer: AI-powered screenshot renaming system using Ollama's Qwen3-VL cloud model
AI-powered screenshot renaming system using Ollama's Qwen3-VL cloud model - skojaku/screenshot-renamer
github.com
October 23, 2025 at 8:31 PM
I've been playing with Ollama Cloud that has a free cloud Qwen 3 Vision Language model, which is surprisingly good. I put together a small app that automatically renames screenshot files based on their content, which has been super helpful for organizing them.
github.com/skojaku/scre...
github.com/skojaku/scre...
Reposted by Sadamori Kojaku
I'd love to see someone try to estimate just how much time and money has gone into research that is either fully undermined by reliance on LLMs or fully pointless --- because obvious if you start from an understanding of what LLMs actually are.
www.pnas.org/doi/10.1073/...
www.pnas.org/doi/10.1073/...
October 18, 2025 at 10:11 AM
I'd love to see someone try to estimate just how much time and money has gone into research that is either fully undermined by reliance on LLMs or fully pointless --- because obvious if you start from an understanding of what LLMs actually are.
www.pnas.org/doi/10.1073/...
www.pnas.org/doi/10.1073/...
Reposted by Sadamori Kojaku
We will have two CoCo seminars on October 29th and 30th! Dakota Murray (U. Albany) and Anastasia Bizyaeva (Cornell) will come to Binghamton/CoCo :) Note the irregular schedule.
October 17, 2025 at 3:08 PM
We will have two CoCo seminars on October 29th and 30th! Dakota Murray (U. Albany) and Anastasia Bizyaeva (Cornell) will come to Binghamton/CoCo :) Note the irregular schedule.
That's a great project to learn how LLMs are trained. $100 is not cheap but the cost to train LLMs comes down to the realm where I can try (just for fun).
github.com/karpathy/nan...
github.com/karpathy/nan...
GitHub - karpathy/nanochat: The best ChatGPT that $100 can buy.
The best ChatGPT that $100 can buy. Contribute to karpathy/nanochat development by creating an account on GitHub.
github.com
October 15, 2025 at 7:09 PM
That's a great project to learn how LLMs are trained. $100 is not cheap but the cost to train LLMs comes down to the realm where I can try (just for fun).
github.com/karpathy/nan...
github.com/karpathy/nan...
Reposted by Sadamori Kojaku
2025 isn't over yet but it's unlikely that Twitter will see a resurgence of research discussions.
October 14, 2025 at 4:27 PM
2025 isn't over yet but it's unlikely that Twitter will see a resurgence of research discussions.
Deep dive into AI podcast
October 8, 2025 at 8:30 PM
Deep dive into AI podcast
Reposted by Sadamori Kojaku
Was your work accepted to #ComplexNetworks2025? It's time to register! Don't wait any longer and take advantage of our early bird discount.🐣
🚨This offer is only available until October 17th.🚨
#ComplexSystems #NetworkScience
Register here:
👉 www.billetweb.fr/complex-netw...
🚨This offer is only available until October 17th.🚨
#ComplexSystems #NetworkScience
Register here:
👉 www.billetweb.fr/complex-netw...
October 6, 2025 at 9:12 PM
Was your work accepted to #ComplexNetworks2025? It's time to register! Don't wait any longer and take advantage of our early bird discount.🐣
🚨This offer is only available until October 17th.🚨
#ComplexSystems #NetworkScience
Register here:
👉 www.billetweb.fr/complex-netw...
🚨This offer is only available until October 17th.🚨
#ComplexSystems #NetworkScience
Register here:
👉 www.billetweb.fr/complex-netw...
Reposted by Sadamori Kojaku
If you’re interested in practicing the philosophy of science in practice why not come along to an online workshop “the philosophy of science in practice - in practice”? #philsci
October 8, 2025 at 1:01 PM
If you’re interested in practicing the philosophy of science in practice why not come along to an online workshop “the philosophy of science in practice - in practice”? #philsci
Reposted by Sadamori Kojaku
This was a fun talk to give. Our new results show that "Meta-learning optimizes predictions of missing links in real-world networks" (arxiv.org/abs/2508.09069). tl;dr: graph neural networks are a limited tool for predicting missing links in most networks
www.youtube.com/watch?v=dRYz...
www.youtube.com/watch?v=dRYz...
GraphGeeks Talk: Meta-Learner That Finds the Best Link Prediction Model
YouTube video by GraphGeeks
www.youtube.com
October 6, 2025 at 2:25 PM
This was a fun talk to give. Our new results show that "Meta-learning optimizes predictions of missing links in real-world networks" (arxiv.org/abs/2508.09069). tl;dr: graph neural networks are a limited tool for predicting missing links in most networks
www.youtube.com/watch?v=dRYz...
www.youtube.com/watch?v=dRYz...
Reposted by Sadamori Kojaku
Network Science, and three of its avengers, featured in the latest @veritasium.bsky.social video (one of the largest youtube channels for Science education)!! youtu.be/CYlon2tvywA?... @stevenstrogatz.com Duncan Watts @barabasi.bsky.social
Something Strange Happens When You Trace How Connected We Are
YouTube video by Veritasium
youtu.be
October 1, 2025 at 6:58 AM
Network Science, and three of its avengers, featured in the latest @veritasium.bsky.social video (one of the largest youtube channels for Science education)!! youtu.be/CYlon2tvywA?... @stevenstrogatz.com Duncan Watts @barabasi.bsky.social
Reposted by Sadamori Kojaku
Come to SUNY and be part of the growing Northeast US complexity community!
Professor and Chair, Department of AI and Society, University at Buffalo
www.ubjobs.buffalo.edu/postings/58816
Professor and Chair, Department of AI and Society, University at Buffalo
www.ubjobs.buffalo.edu/postings/58816
Professor, Department Chair of AI and Society
The Department of AI and Society at the University at Buffalo, invites applications from outstanding scholars with excellent leadership abilities for the position of Department Chair. The Department o...
www.ubjobs.buffalo.edu
October 3, 2025 at 6:21 PM
Come to SUNY and be part of the growing Northeast US complexity community!
Professor and Chair, Department of AI and Society, University at Buffalo
www.ubjobs.buffalo.edu/postings/58816
Professor and Chair, Department of AI and Society, University at Buffalo
www.ubjobs.buffalo.edu/postings/58816
One of the time-consuming tasks in paper writing is to curate bibtex for references. So I automated bibtex curation with an Alfred app that fetches entries from dx.doi.org using DOI or title. It saves time, reduces errors, and maintains consistent LaTeX keys. github.com/skojaku/tobi...
GitHub - skojaku/tobibtex
Contribute to skojaku/tobibtex development by creating an account on GitHub.
github.com
October 2, 2025 at 3:41 PM
One of the time-consuming tasks in paper writing is to curate bibtex for references. So I automated bibtex curation with an Alfred app that fetches entries from dx.doi.org using DOI or title. It saves time, reduces errors, and maintains consistent LaTeX keys. github.com/skojaku/tobi...