iiinova — where your digital self keeps evolving

Frank Dellaert @fdellaert.bsky.social · 02/06/2026

Proud to have contributed a tiny part to this cool paper.

Reposted by Frank Dellaert

Peter Gleick @petergleick.bsky.social · 31/05/2025

There are 2 previous historical cases of countries destroying their science and universities, crippling them for decades: Lysenkoism in the USSR and Nazi Germany. The Trump administration will be the 3rd. It's not just budgets but research, institutions, expertise, and training the next generation.

Graphs showing 25 years of budgets for the National Institute of Health, NASA, and the NSF. In all cases, the proposed budget for next year is far, far below any year of the previous quarter century.

449151517827

Reposted by Frank Dellaert

Kosta Derpanis @csprofkgd.bsky.social · 04/05/2025

I know this look well. It’s the same one I get when I introduce the Fourier transform in my computer vision course to the CS crowd.

1221

Reposted by Frank Dellaert

The Onion @theonion.com · 31/03/2025

Trump Says He Won’t Rule Out Third Reich

261130462507

Frank Dellaert @fdellaert.bsky.social · 31/03/2025

Part 2 of SLAM handbook is out for public comments! let us know what you think :-) Issue tracker on GitHub awaits! Link: github.com/SLAM-Handboo...

04911

Reposted by Frank Dellaert

George Takei @georgetakei.bsky.social · 30/03/2025

And yet Joe Rogan puts guests on who say "Vaccines aren't actually responsible for the reduction in infectious diseases.”

The image is a heatmap titled "Measles," showing measles cases across U.S. states from 1928 to 2003. States listed on the y-axis include Alaska, Arkansas, California, Connecticut, Delaware, Georgia, Iowa, Illinois, Kansas, Louisiana, Maryland, Michigan, Missouri, Montana, North Dakota, New Hampshire, New York, Oklahoma, Pennsylvania, South Carolina, Tennessee, Utah, Vermont, Wisconsin, and Wyoming. The x-axis spans years from 1928 to 2003. A vertical line marks the introduction of the measles vaccine around 1963. The color scale at the bottom ranges from blue (0 cases) to red (4,000+ cases), with shades of green, yellow, and orange indicating intermediate values (1,000, 2,000, 3,000 cases). Before 1963, many states show frequent high case numbers (yellow to red), while after 1963, cases drop significantly, with mostly blue indicating near-zero cases.

903230116742

Reposted by Frank Dellaert

Andrew Davison @ajdavison.bsky.social · 28/02/2025

Aria Gen 2 is very impressive, with fully onboard SLAM and various other perception all within a 75g device with an hours-long battery life. Processing on custom silicon. Congrats to the Reality Labs team.

meta.com

Introducing Aria Gen 2: Unlocking New Research in Machine Perception, Contextual AI, Robotics, and More

Today, we're excited to announce the next step in the Project Aria journey: the introduction of Aria Gen 2 research glasses.

1243

Reposted by Frank Dellaert

Riku Murai @rmurai0610.bsky.social · 25/02/2025

MASt3R-SLAM code release! github.com/rmurai0610/M... Try it out on videos or with a live camera Work with @ericdexheimer.bsky.social*, @ajdavison.bsky.social (*Equal Contribution)

25110

Frank Dellaert @fdellaert.bsky.social · 18/02/2025

Some personal news: as of January I am back full-time at Georgia Tech following a 2-year leave as Verdant Robotics’ CTO. I will continue to be involved with Verdant as part-time Chief AI Officer, thinking strategically about the role of AI in Robotics for Ag.

0290

Frank Dellaert @fdellaert.bsky.social · 12/02/2025

Gemini is good but too verbose :-)

030

Frank Dellaert @fdellaert.bsky.social · 09/02/2025

The visual system of a jumping spider is fascinating. Look at those cones behind the fixed main lenses! The retinas are at the end of the cones. youtu.be/gvN_ex95IcE?...

youtu.be

Yellow amycine jumping spider, Reserva Canadé, Ecuador

YouTube video by wmaddisn

181

Reposted by Frank Dellaert

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 06/02/2025

We've built a simulated driving agent that we trained on 1.6 billion km of driving with no human data. It is SOTA on every planning benchmark we tried. In self-play, it goes 20 years between collisions.

2229855

Reposted by Frank Dellaert

Alexandre Morgand, PhD @alexmrgd.bsky.social · 22/01/2025

Video Depth Anything: Consistent Depth Estimation for Super-Long Videos TL;DR: Long videos support; Depth Anything V2 with efficient spatial-temporal head. Temporal consistency loss -> depth gradient (no geometric priors)

2297

Frank Dellaert @fdellaert.bsky.social · 21/01/2025

Maybe a broom attached to the train :-)

040

Frank Dellaert @fdellaert.bsky.social · 21/01/2025

That’s such a cool idea! (Thanks K for flagging this to me)

190

Frank Dellaert @fdellaert.bsky.social · 12/01/2025

Not sure I’m optimistic about that :-) I can see the number of humans involved decrease. Georgia Techs OMSCS is a good metaphor. More TAs, less teachers. TAs will be replaced by AI: one AI-TA per student per course.

000

Frank Dellaert @fdellaert.bsky.social · 12/01/2025

I’m sure oxbridge tutors have a pretty good idea of what they have in front of them after a couple of sessions :-)

000

Frank Dellaert @fdellaert.bsky.social · 12/01/2025

I don’t have the answers. But I’m optimistic we’ll arrive at a good outcome eventually. We will still need to find a way to somehow contribute value, and to document that fact in order to get hired for those skills. Maybe the key to certification lies in those 1-1 tutoring sessions themselves.

100

Frank Dellaert @fdellaert.bsky.social · 12/01/2025

Sure. I think we’ll have to find a different way. At this point when I code, I use ChatGPT/Claude extensively, so maybe we should start thinking about this being a valid way to complete assignments :-)

200

Frank Dellaert @fdellaert.bsky.social · 12/01/2025

I am more optimistic. It’s well known that one on one tutoring can be up to two standard deviations more effective than classroom teaching, but it’s prohibitively expensive. AI gives us cost-effective solution for just that. It’s *already* doing that for me :-)

100

Frank Dellaert @fdellaert.bsky.social · 08/01/2025

First iRIM@GT seminar of 2025: Ani Majumdar! Well attended!

040

Reposted by Frank Dellaert

Jeff Dean @jeffdean.bsky.social · 05/01/2025

Getting myself set up here. I found the Sky Follower Bridge Chrome plugin pretty helpful (thanks @kawamataryo.bsky.social!) chromewebstore.google.com/detail/sky-f...

chromewebstore.google.com

Sky Follower Bridge - Chrome Web Store

Easily transfer your following users and list members from X to Bluesky.

913818

Frank Dellaert @fdellaert.bsky.social · 04/01/2025

3D feed-forward Gaussian Splatting feels like magic, let alone 4D!

0160

Reposted by Frank Dellaert

Robert "The Bobby Yaga" McNees @mcnees.bsky.social · 01/01/2025

100 years ago today, #OTD in 1925, Edwin Hubble announced that Andromeda and other spiral nebulae were definitely separate galaxies outside the Milky Way, in a paper read to an AAS meeting by H.N. Russell. There was no doubt that the Universe was more than just our little island of stars. 🧪 🔭 ⚛️

A black and white photo of M31, the Andromeda galaxy. It is characterized by spiral arms and dark lanes of dust and gas, with a bright center.

423583540

Frank Dellaert @fdellaert.bsky.social · 18/12/2024

Yes, like copilot. But then copilot is not as smart as 4o let alone o1.

000

Frank Dellaert @fdellaert.bsky.social · 18/12/2024

ChatGPT canvas is cool. Maybe a bit slow still.

110

Reposted by Frank Dellaert

Alfredo Canziani @alfcnz.bsky.social · 13/12/2024

Convolutional Differentiable Logic Gate Networks @FHKPetersen

3453

Frank Dellaert @fdellaert.bsky.social · 14/12/2024

Today I used ChatGPT canvas to help me simplify the SE_2(3) exponential map calculation, and its Jacobian. 4o looks dumb in comparison to o1 now, though :-(

040

Reposted by Frank Dellaert

linyijin.bsky.social @linyijin.bsky.social · 13/12/2024

Introducing 👀Stereo4D👀 A method for mining 4D from internet stereo videos. It enables large-scale, high-quality, dynamic, *metric* 3D reconstructions, with camera poses and long-term 3D motion trajectories. We used Stereo4D to make a dataset of over 100k real-world 4D scenes.

25912

Frank Dellaert @fdellaert.bsky.social · 12/12/2024

Ran out of quota, switching to Gemini 2.0 Flash. It is very good as well. And he has way more context than 4o, which was my first backup.

010

Frank Dellaert @fdellaert.bsky.social · 12/12/2024

Back in college I took a developmental biology class. This field has come so far with advancing technologies like light sheet microscopy. This video is just plain amazing.

070

Frank Dellaert @fdellaert.bsky.social · 12/12/2024

I’m talking differential geometry with ChatGPT o1 like there is no tomorrow. This thing is amazing. On the other hand, I might need my own nuclear power station.

3200

Reposted by Frank Dellaert

Noah Snavely @snavely.bsky.social · 06/12/2024

Zhengqi led a really nice new paper that computes really nice camera poses and depth maps from everyday videos, not just the standard SLAM-mable sorts of videos you often see. It's really fast and robust, and I think it's quite neat.

1334

Reposted by Frank Dellaert

Andrew Davison @ajdavison.bsky.social · 06/12/2024

Fit-NGP: millimetre accurate 3D object pose estimation from RGB images only via Instant NGP. What can we do in manipulation with this level of accuracy? From Marwan and Ignacio from the Dyson Robotics Lab, Imperial College, ICRA24. marwan99.github.io/Fit-NGP/ youtu.be/KQ7yH_em3Qg?...

youtu.be

[ICRA 2024] Fit-NGP: Fitting Object Models to Neural Graphics Primitives

YouTube video by Dyson Robotics Laboratory at Imperial College

1242

Frank Dellaert @fdellaert.bsky.social · 04/12/2024

Thanks! I’ll pass this on to our latex guru :-)

120

Reposted by Frank Dellaert

Chris Paxton @cpaxton.bsky.social · 03/12/2024

I'd like to introduce what I've been working at @hellorobot.bsky.social: Stretch AI, a set of open-source tools for language-guided autonomy, exploration, navigation, and learning from demonstration. Check it out: github.com/hello-robot/... Thread ->

613223

Reposted by Frank Dellaert

ruiqigao.bsky.social @ruiqigao.bsky.social · 02/12/2024

A common question nowadays: Which is better, diffusion or flow matching? 🤔 Our answer: They’re two sides of the same coin. We wrote a blog post to show how diffusion models and Gaussian flow matching are equivalent. That’s great: It means you can use them interchangeably.

625459

Frank Dellaert @fdellaert.bsky.social · 03/12/2024

I have been to Rottnest Island and quokka fur slippers make sense to me :-) en.m.wikipedia.org/wiki/Rottnes...

en.m.wikipedia.org

Rottnest Island - Wikipedia

010

Reposted by Frank Dellaert

Michael J. Black @michael-j-black.bsky.social · 03/12/2024

MoSh has won the 2024 SIGGRAPH Asia Test-of-Time Award. What’s MoSh? It takes motion capture markers and returns the animation of a realistic 3D human body in #SMPL-X format. I wrote a blog post to explain why MoSh is still relevant after 10 years. perceiving-systems.blog/en/news/moti...

45412