
CV4PhysicalAI 🤖 PhD @Ukita_Lab, DC1, Toyoaki, Tech chair @ACCVConf @liveinvista, Dad🇻🇳 Ex: Intern @Woven_ToyotaJP, @ToyotaCRDL, @TTIC_Connect, TA @Matsuo_Lab
Toyota Technological Institute
Joined February 2021
- Tweets1.5K
- Following740
- Followers403
- Likes21.9K
Pinned Tweet
📢 Call for Papers @ACCVConf !📢
The 1st HUFPEI Workshop at #ACCV2026 is open for submissions!
📝 Track 1: New papers, archival. Due Oct 1
🗣️ Track 2: Papers published in 2026 (including ACCV main papers), presentation-only. Due Oct 22
📍 Dec 15, Osaka
hufpei-workshop.github.io/ac…
Hiromu Taketsugu retweeted
ACCV 2026にて、12/15午前に名工大の丁先生・玉木先生と一緒にスポーツ映像理解のチュートリアルを行います!
選手の検出・追跡から言語モデル・戦術理解、一般的な動画理解まで紹介します。
ning-d.github.io/accv2026-sp…
Hiromu Taketsugu retweeted
Kotoba Technologies is joining ElevenLabs!
Kotoba Technologiesが@ElevenLabs に参画することをご報告いたします。
これまでKotobaを支えてくださった皆さまに、心より感謝申し上げます。
Kotobaの第2章が、ここから始まります。
東京から世界へ、次世代の音声AIを届けていきます。
Hiromu Taketsugu retweeted
🎉 Our paper "Hexa-P: Physical Plausibility-aware Pre-training for Pedestrian Path Prediction" has been accepted to Pattern Analysis and Applications!
🤖 We pre-train on large-scale synthetic pose-path data from a physics simulator, then transfer to real multi-agent scenes 🚶
Hiromu Taketsugu retweeted
I'm co-organizing the 3rd Spotting Birds challenge at #MVA2027 🐦
New this year: forecast each bird's next 5 s from 2 s of raw drone video, not past trajectories, so a bird your tracker misses can't be forecast.
And 41% of birds drop out of view at some point in those 5 s.
📣 Call for participation: Small Multi-Object Tracking and Prediction for Spotting Birds (SMOTP4SB) Challenge at #MVA2027
Can your model track birds only a few pixels wide in drone video, and predict where each one will fly in the next 5 seconds?
🧵👇
Hiromu Taketsugu retweeted
🔮 Track 2 (new): Bird Trajectory Prediction
Given 2 seconds (60 frames) of raw video, track the birds and predict each one's position over the next 5 seconds (150 frames).
Metric: the new SO-HOTA-TP, which also penalizes predictions after a bird leaves the view.
Hiromu Taketsugu retweeted
📣 Call for participation: Small Multi-Object Tracking and Prediction for Spotting Birds (SMOTP4SB) Challenge at #MVA2027
Can your model track birds only a few pixels wide in drone video, and predict where each one will fly in the next 5 seconds?
🧵👇
Hiromu Taketsugu retweeted
At #CoRL2026 this week 🤖 Say hi if you see me around!
Hexa-P also contributes to safer crowd navigation!
🎉 Our paper "Hexa-P: Physical Plausibility-aware Pre-training for Pedestrian Path Prediction" has been accepted to Pattern Analysis and Applications!
🤖 We pre-train on large-scale synthetic pose-path data from a physics simulator, then transfer to real multi-agent scenes 🚶
🌐 Project page: iminthemiddle.github.io/Hexa…
📑 Springer (Paper): link.springer.com/article/10…
🐙 GitHub (Code): github.com/ImIntheMiddle/Hex…
🤗 Hugging Face (Weights&Data): huggingface.co/ImIntheMiddle…
Today, we’re announcing the next phase of our autonomous driving collaboration with @ToyotaMotorCorp to scale a fleet of bZ4X autonomous vehicles in Japan, beginning with testing and validation in 2027.
Read more about our partnership here: waymo.com/blog/2026/10/toyot…
__
本日、@ToyotaMotorCorpとの自動運転分野における協業を新たな段階へと進めることを発表します。2027年に開始する試験および検証を皮切りに、日本国内で「bZ4X」の自動運転フリートの拡大を進めてまいります。
パートナーシップの詳細はこちらをご覧ください:
waymo.com/blog/2026/10/toyot…
Our paper, “Hexa-P: Physical Plausibility-aware Pre-training for Pedestrian Path Prediction,” has been accepted for publication in Pattern Analysis and Applications! 🎉
Huge thanks to my co-authors, Prof. Nobuhara and Prof. Ukita 🙏
More details coming soon!
11
372
Hiromu Taketsugu retweeted
This is a virtual human body with 700 muscle actuators, 206 joints, and 90 body segments.
It's fully open source, and researchers can use it to simulate human movement through muscle-level control.
Hiromu Taketsugu retweeted
🎉 Excited to share that our paper “Where Did the Light Come From? Diagnosing Cast-Shadow Reasoning in VLMs and Borrowing the Missing Shadow Prior from Image Generation Models” has been accepted to #WACV2027 !!
arxiv.org/abs/2610.30740
Huge thanks to all collaborators☀️👤
Hiromu Taketsugu retweeted
“The more you delegate, the more your fundamentals matter, because you cannot supervise what you do not understand.”
Yesterday a junior researcher messaged me, frustrated and a little scared. They had just read that a company pretrained its model with a single researcher. One person for the entire pretraining. no R&D team, no nothing.
If one person and a cluster can do what used to take a whole team, then the expertise they had spent years building is already obsolete.
I get the fear. Marketing headlines like this are built to produce it. But I think the conclusion is wrong, and the reasons it is wrong say a lot about what a researcher is for now. Here is what I told them, expanded:
1. Be adaptive: RSI is real, but this isn't it yet
Recursive self-improvement is real, and you should plan around it. In September, OpenAI said its research org now logs 3.1 agent-workdays per human workday, and it aims for an automated AI researcher by March 2028. Models may soon run most of pre- and post-training. If your five-year plan assumes otherwise, change the plan.
We have already seen lots of open-source releases from the major players, with open weights and detailed technical reports. With that much public know-how, it is not hard to copy a mundane architecture (probably with sliding-window attention) and train something, as long as you have the compute it requires.
That makes it an execution and budget problem, not a research problem.
If one person can train a model by following recipes that are already public, it shows those recipes have matured. It does not show that the people who invented them are no longer needed.
What stays scarce is the compute to run those recipes and the ideas about what to try next. The rest of my tweet is about the ideas.
2. The role grows beyond repeating what's known and safe
If an agent and a cluster can reproduce what is already known, then reproducing what is already known stops being a career. If most of your job is rerunning last year's recipe with small tweaks, that job is shrinking fast. That is the uncomfortable half of the message.
The other half is that auto-research does not work yet, at least not in the sense that matters. It does work, and impressively, where someone has already defined the problem and built the scoreboard. On the EinsteinArena [1] platform, agents set new records on 11 open math problems, including raising the best known lower bound for the kissing number in dimension 11 from 593 to 604.
Notice that every one of those problems was chosen because a computer could score the answers automatically. Humans picked the problems and wrote the scoring code and the agents did the searching.
Then there is the fully AI-generated paper [2] that got through peer review. It sounds big until you read the details. It passed only the first round (was later withdrawn), at a workshop rather than the main conference, and that workshop accepts 70% of submissions.
None of this is meant to discredit agents. The point is that we complement each other. Agents are great at the work that used to take many weeks. Basically, writing code, running experiments, debugging pipelines. We bring what they still lack which is the idea worth testing, and the judgment to tell a real result from a lucky one.
Even inside a lab as agent-heavy as OpenAI, humans hold the top of the stack. Based on their own report, "People still set our research priorities, judge which ideas and results to pursue" [3]. High-level planning is still a tiny share of what its agents produce, and OpenAI expects the least automatable work to take a growing share of researcher effort. That is exactly where you want to be standing.
Combine an agent’s speed with your own ideas and judgment, and you get the 10x researcher and juniors can be one too.
3. Judgment becomes the job, so fundamentals matter more
When code, experiments and plausible-looking results cost almost nothing, the scarce skill is judgment. Which experiment to run, which result to believe, which direction to drop. That is the part the one-person pretraining run left to the human. It is also the hardest part to fake.
Agents produce confident output very fast. They will happily report a three-point gain that came from a leaky eval, a data-loader bug, a lucky seed or a tokenizer mismatch. Then they will write the ablation table and the story to match.
You can see the need for oversight even in the success stories. The EinsteinArena team had to keep auditing and hardening its verifiers as agents optimized aggressively against them. At OpenAI, over half of the successful agent tasks that would take a human 4–8 hours still involved at least one human intervention [3].
The person who catches the bad result is the one who understands the system. Why the loss curve has that shape, what a realistic gain looks like, what a too-good result looks like. That is the paradox of this era. The more you delegate, the more your fundamentals matter, because you cannot supervise what you do not understand.
So build the foundations on purpose. Implement attention by hand at least once, then a linear variant like Gated DeltaNet. Train a small model end to end, then break it deliberately and watch what happens. Read papers critically, appendix included, and learn what the numbers should look like before you see them.
That knowledge is what lets you direct the machine instead of simply believing it.
4. LLMs are far from solved
The one-researcher model was possible because the method was already public. Your edge is in the questions nobody has answered yet, and you don't need to leave LLMs to find them:
(1) How do we learn more from less data?
(2) Can linear attention like Gated DeltaNet match full attention on retrieval and long context, and what is the right mix of linear and full attention layers in a hybrid?
(3) How can a model keep learning after training without forgetting what it knew?
(4) Why does reasoning get less reliable over long tasks, and how do we fix it?
(5) Why do models make things up, and can they learn to say "I don't know"?
(6) How do we measure progress when benchmarks saturate or leak into training data?
(7) What is actually happening inside these models, and how do we keep increasingly autonomous ones doing what we intend?
Recursive self-improvement makes these questions more urgent. If models are going to train models, someone has to decide what they optimize for, how we measure it, and how we keep the loop from drifting.
RSI is a research agenda, not the end of one.
So, are you obsolete?
No. One version of the job is the one where research meant reproducing what was already known and safe. That version was never the point.
Learn your fundamentals deeply enough to judge the machine, and your agentic stack well enough to make it work for you. Then spend that leverage on ideas nobody has had yet. The machines are getting very good at driving on roads we have already paved; someone still has to decide where the next one goes.
[1]: arxiv.org/abs/2606.10402
[2]: arxiv.org/abs/2606.15497
[3]: openai.com/index/research-ac…
Hiromu Taketsugu retweeted
If you use SAM 3D Body for digital humans, robotics, or video generation, we have a cool new plugin for you: CLUE!
CLUE is an easy-to-use plugin (3 lines!) to make SAM 3D Body probabilistic, providing per-vertex uncertainty!
Details, Hugging Face Demo and other stuff in🧵
Hiromu Taketsugu retweeted
"How can we navigate the growing field of Physical AI safety?"
🚀 Introducing 𝗔𝘄𝗲𝘀𝗼𝗺𝗲 𝗣𝗵𝘆𝘀𝗶𝗰𝗮𝗹 𝗔𝗜 𝗦𝗮𝗳𝗲𝘁𝘆 — an interactive, daily-updated research catalog!
✅ 6-axis taxonomy
✅ Interactive cross-maps
✅ LLM-assisted curation
🌐 futakw.github.io/Awesome-Phy…
🤖 Made with AI
Hiromu Taketsugu retweeted
AIの日本語を推敲するスキル『yomiyasu (よみやす) 』
v2.0.1 をリリースしました!🎉
アルゴリズムの最適化を磨きこみ、計算量を約 O(n³) から O(n) まで落とすことに成功しました。
37万字の文章で6分12秒かかっていた検査が 1.4秒 で終わります。
出てくる結果は v2.0.0 と1文字も変わりません!
AI臭くて読みにくい文章を自然で読みやすい日本語へ書き直すスキル『yomiyasu』をリリースしました。
従来の「単語を禁止する手法」から改善し、統語構造(誰が何をどうした)の復元と、比喩動詞の具体化に特化しています。
Claude Code, Codexなどで今日から利用できます。
github.com/nanaism/yomiyasu
Hiromu Taketsugu retweeted
こちら,2nd trackまだまだ募集中ですのでぜひよろしくお願いいたします!
📢 Have a paper accepted in 2026? Present it at HUFPEI @ACCVConf #ACCV2026! 📢
✅ Any conference or journal, including ACCV main
✅ No review, a relevance check only
✅ Non-archival
⏰ Apply by Oct 22: just send title and authors!
📍 Dec 15, Osaka
hufpei-workshop.github.io/ac…





















