Justin Solomon
@JustinMSolomon
Followers
7K
Following
17K
Media
264
Statuses
2K
Associate prof, MIT EECS/CSAIL 💻🔬🦥🧀🏋️♀️🎼🐈🎻🏳️🌈 he/him/his — justinmsolomon@🟦🌤️🦋📲
Somerville, MA
Joined October 2011
Just wrapped up filming the last lecture for my graduate "Shape Analysis" course (@MIT @MITEECS @MIT_CSAIL 6.838)! Check out this playlist for 36 hours, 57 minutes, 39 sec of [admittedly idiosyncratic] lectures covering topics in applied geometry: 🎉🚨🔥💻.
5
52
348
Randomly came across this post — finally someone gets me! Thanks, random internet student!
8
19
769
Want to learn about computer graphics? Feeling lonely during quarantine? In either case, check out my Intro to Computer Graphics course, where you can hear me talk about graphics nonstop for 26 hours, 33 minutes, and 11 seconds (professionally captioned!).
11
117
511
Announcing SGI 2022! Undergrads and MS students: . Apply for 6 weeks of paid summer geometry processing research. No experience needed: 1 week tutorials + 5 weeks of projects. Mentors are top researchers in this emerging branch of graphics/computing/math.
13
129
341
Gave a talk and used the phrase “classical theorem by an old dead mathematician.” Turns out he was in the room.
9
20
315
Announcing SGI 2025! Undergrads and MS students: Apply for 6 weeks of paid summer geometry processing research. No experience needed: 1 week tutorials + 5 weeks of projects. Mentors are top researchers in this emerging branch of graphics/computing/math.
4
54
307
That fateful moment as a new PhD student when I unwittingly chose the PowerPoint template I would use for the rest of my academic career.
7
2
280
Announcing SGI 2024! Undergrads and MS students: Apply for 6 weeks of paid summer geometry processing research. No experience needed: 1 week tutorials + 5 weeks of projects. Mentors are top researchers in this emerging branch of graphics/computing/math.
7
78
267
My group will be seeking new PhD students in the coming cycle! The best way to reach us is to apply to the @MITEECS PhD program. I look for independent, creative, interactive, supportive, passionate, bright students who want to work on fundamental problems with a geometry flavor.
4
39
254
Announcing SGI 2023! Undergrads and MS students: Apply for 6 weeks of paid summer geometry processing research. No experience needed: 1 week tutorials + 5 weeks of projects. Mentors are top researchers in this emerging branch of graphics/computing/math.
3
82
252
Just got rejected from a grant I didn’t apply for. I guess that’s more efficient than the usual rejection where you write a proposal first?.
9
7
218
Obligatory post-tenure news article 🗞️🥳 -- plus a very purple photograph!.
17
4
167
Wasserstein barycenters are useful in stats/ML but typical algorithms discretize the domain, leading to low-dimensional/coarse approximations. With PhD student Lingxiao Li, postdoc Aude Genevay @a_gnv, and @MITIBMLab's Misha Yurochkin we construct a continuous barycenter!🔥🎊🎈🎉
Continuous Regularized Wasserstein Barycenters. (arXiv:2008.12534v1 [cs.LG])
0
34
161
ML+vision Twitter: I know nothing about "equivariant" NNs (i.e. NNs satisfying f(RX)=R'f(X) for rotations/rigid motions R and 3xN point clouds X). There's a giant cloud of papers that overlap in confusing ways, and the surveys seem to be piles of citations. What should I read?.
13
14
161
Starting another semester of disembodied online teaching from video. This time a grad level applied geometry course — publicly visible lecture videos coming soon! Experimenting with a “LightBoard” MIT made available for recording.
5
6
160
Each time I teach differential geometry, I dread covering why we need covariant, Lie, and exterior derivatives, which all have similar (but not identical) input/output. I’m pleased to announce this year I’ve done an equally poor job. But this time it’s on a stylish glass board.
3
5
157
48 of 75 @NeurIPSConf reviews (due tomorrow) I'm in charge of as AC are missing, and many I've received are not thorough to say the least. C'mon folks, show some investment in your community and be careful/prompt. You're the same people submitting all these papers to begin with!.
7
12
148
Attention undergrads and MS students: Join us this summer for a 6-week paid geometry processing research program! No experience needed: One week tutorials, then 5 weeks projects. Work with top faculty in this emerging branch of graphics/computing/math.
6
41
145
Being able to stand behind the blackboard is allowing me to pioneer new frontiers of mathematical messiness not previously known to be possible!
6
4
140
Been trying to learn Jax. So far, I find it an amazing, inspirational piece of software and also have a splitting headache and want to throw my laptop out the window.
4
4
145
Share your favorite topic in geometry/optimization/numerics/ML/baking/cello and I can try my best to explain it at the board!.
Ask us your questions about geometry, computer graphics, and ML. We’ll pick one to feature in an upcoming explainer w/MIT professor Justin Solomon (@JustinMSolomon). For more on his work:
11
16
141
Not sure why it bothers me so much, but can we stop using the word “traditional” in academic papers to refer to methods that came out 3 years ago?.
10
4
137
YouTube comments are usually awful but sometimes they make your day 😭
11
6
131
Excited to share that the 2024 Symposium on Geometry Processing (SGP) @GeometryProcess will be at @MIT! This will be the first (non-virtual) SGP to be in North America. SGP is a premier venue for research in applied geometry, including fabrication, VR, 3D vision, & architecture.
3
13
132
🚨 Seeking a postdoc starting 2025 (flexible timing) to participate in a collaboration between my group at @MIT_CSAIL and the @MITIBMLab (PI: @Yurochkin_M) focusing on the intersection of foundation models and geometric data processing.
2
26
130
Missed SGP 2024 and want to catch up on the latest in geometry processing research in HD? Want to hear tutorials from leading researchers in the field? Attended SGP and want to relive the magic? Check out the official conference Youtube channel:
1
30
125
A few years back I had a question on MathOverflow closed as trivial and had to gently explain I was a postdoc in a fancypants math department seeking help with an open research problem. Resolving the question took years and led to a nice paper. The internet is a weird place.
6
18
122
2024 will be a big year for geometry at @MIT_CSAIL/@MITEECS: We'll be organizing/hosting BOTH the Symposium on Geometry Processing (SGP) and Summer Geometry Initiative (SGI)!. We are excited to bring geometry to @MIT, but it will come at substantial cost---so we need your help!
1
19
121
Thank you to the authors for a detailed response that addresses all my concerns and provides the experiments I requested. I have no further questions or comments and choose to leave my score the same.
6
7
113
Practice run for teaching graphics class this fall. Toilet paper camera stand works remarkably well. 2020 amirite? 🧻
7
2
112
🎩🥳.
🎩 A tip of the cap to Prof. Justin Solomon (@JustinMSolomon) of MIT for winning the ACM SIGGRAPH Significant New Researcher Award for his research focusing on geometric problems in graphics and machine learning. Kudos to you 👏👏.
4
1
107
Just finished 2 weeks of COVID+strep (hit the jackpot!). Symptoms were relatively mild, although I wouldn’t recommend it! But, this helped me observe some habits in academia that aren’t helpful during a pandemic (I’m guilty of many). Disorganized notes below. 1/n.
10
5
109
Had some of my first in-person research discussions with students and collaborators in 1.5 years. Maybe it’s just the novelty, but the take away seems to be that tablets, Zoom, Slack, and other digital tools don’t come close to writing equations together on a shared whiteboard.
1
2
106
Tomorrow, we kick off the Summer Geometry Institute (SGI) 2021, an experiment in online collaboration and education engaging students from across the world in applied geometry research.
2
18
106
Today, our brilliant student Lingxiao Li presented "Interactive All-Hex Meshing via Cuboid Decomposition" at @SIGGRAPHAsia. It provides a UI+clever geometric algorithms for interactive polycube-based hex meshing. This is the first time I've been able to hex mesh reliably!📦🥳
2
19
103
Someone posted these all over campus. Certainly an interesting contrast with the current (top-down) policy decisions at @MIT …
3
13
100
Does your company want to promote and support the next generation of researchers in geometry, computer vision, graphics, robotics, and machine learning? Consider sponsoring the Summer Geometry Initiative (SGI), entering its third year in 2023!
2
26
98
On the whole, I've been really impressed with @YouTube's auto-captioning technology --- with one notable exception.
3
1
94
That time of year where I wonder why we all stake our careers on a glorified random number generator whose output we regard as a respected mark of quality. \end{rant}.
The feeling when you read a @NeurIPSConf review of your paper and find it equaly well-informed as the opinion of a random person you meet in a shopping mall. .
2
9
97
Our group in @MIT_CSAIL will likely (finalizing funding) have an opening for a postdoc at the intersection of ML and geometric representations (eg meshes, CAD, LiDAR) to start late 2020/early 2021. Check out our web page to get an idea of what we’re up to:
3
19
90
Unexpected application of optimal transport: Our intern/@MIT undergrad (now MEng) Trevor used it to design audio transitions, creating a "generalized portamento" that matches and slides between pitches: . All credit to Trevor for this cool idea!.
2
12
89
Superstar undergrad Zoë Marschner and superstar PhD students Paul Zhang+David Palmer are at it again! ⭐🌟 "Sum-of-Squares Geometry Processing" (SIGAsia21) gives a new tool in the toolbox for algebraic problems involving higher-order primitives.
2
16
86
Annual grumble reading PhD apps:.1. Focus your rec letter on the student, not your CV. 2. Don't tell us they're pleasant/caring. Do you say that about white cis men?.3. Don't qualify ("One of the best women").4. Don't address the letter to Sir/Gentlemen.Cmon, it's 2023 folks.
4
6
86
Since COVID keeps canceling my sabbatical plans, I'm giving up & doing a "staybbatical" (?) this spring. Seems like a good time to revise Numerical Algorithms book for a 2nd ed. What topics would y'all suggest adding/revising for a modern applied numerical methods/modeling text?.
5
4
80
Big week for students in our @MIT_CSAIL Geometric Data Processing group—superstar undergrad Zoë Marschner got the @MITEngineering-wide Outstanding UROP Award and PhD students Lingxiao Li+David Palmer got the @MITEECS Hennie Teaching Award for the past fall+spring (resp) 🎉🥳🎈🎊.
2
0
80
Challenged @MIT_CSAIL to promote our students’ more theoretical non-ML work. Adam did a great job explaining fairly abstract research on frame fields and meshing! Check out these papers at @GeometryProcess and @siggraph.
0
10
79
To whom it may concern:. I am writing on behalf of [name] in support of their graduate school application. Let me introduce myself. [5 paragraphs with resume, TR35 awards, grants, SAT score, max deadlift]. [Name] worked in my lab. Sincerely, [prof].Named professorship, [school].
2
1
78
I think I’m done giving talks/visits to industrial organizations that offer no compensation/funding upfront. Had several experiences lately with companies that ask a lot of my/my students’ time with vague promise of support/collaboration and then ghost when I follow up. </rant>.
6
0
79
I’ve been shocked by the petty, childish behavior @zicokolter is calling out. If you have an issue, take it up with the authors or write your own article with a counterpoint. Don’t proclaim that a student doesn’t deserve an award because you think you found a hole in their work.
But we need to be honest about who bears 99% of the brunt of this very public and immediate criticism: the PhD student first authors. I can't *fathom* how stressful this must be for them. *I* have been stressed just thinking about it, and I have zero connection to the papers. 4/N.
4
3
78
Annual reminder that the "about me" section of your recommendation letter probably shouldn't be longer than the part where you write about the student.
4
2
75
I too am tired of meaningless questions like this on rec forms. How should I estimate this %? Have I recently recommended over 100 students to assess 1-2% vs 5%? Do you want a confidence interval? And as a reader of admissions files, apparently 99% of students are in the top 1%.
5
1
74
The Summer Geometry Initiative (SGI) finally got approved by MIT to continue in 2022 (yay!), but to proceed we need to raise funds to pay the 2022 Fellows' stipends. If you work in industry or have contacts who might wish to sponsor SGI, please contact me!.
1
22
73
I’ve noticed a huge uptick in postdoc/PhD/SGI applicants who are suddenly “keen” to do everything.
ChatGPT annoyingly tries to insert "delve into" in every paragraph! I am immediately suspicious when I see students' work with this term. Are there any other tell-tale phrases? Here is the Google Trends graph for "delve into":
3
5
72
I don’t think I’ve ever been live-tweeted before! What an honor 🏆 — thanks @ilkedemir 🙏.
Our last keynote of the day: Geometry Processing-Inspired Deep Learning, which also happens to be @JustinMSolomon's first time at ICCV? 😲 #DLGC2021 #ICCV2021
0
4
72
Can you hear the shape of a book chapter? Probably not! But you *can* learn all about this and other spectral geometry problems in our new survey/book chapter: —a serious undertaking with student Yu Wang to catalog and explain where these problems appear.
1
13
71
Chrome just added a feature where tabs can be grouped and hidden from view, sealing the fate of the many papers I’ve had open for months as tabs to read some day.
3
0
67
🚨🚨 Seeking project leaders & volunteers for the Summer Geometry Initiative (SGI) 2022! Join this summer to mentor the next generation of researchers in geometry, graphics, vision, and related fields. Flexible timing/commitment, 100% online, 200% rewarding. Links below.
1
17
64
SGI mugs are here! 🤩🤩 Design created by our awesome SGI Fellows. Next week’s project: assembling packages of swag to mail to the students and mentors.
1
1
66
Receiving sponsored/targeted Twitter ads about colleagues’ CVPR papers and am not sure how I feel about this.
4
5
65
And finally @rohansawhney1 and Bailey Miller wrap up the 2024 SGP Graduate School by introducing Monte Carlo Geometry Processing! 🎉
2
3
65
Main conference day 2 to an energetic start with @_AlecJacobson ! Got a preview during the AV setup this morning and am excited to hear his most controversial opinions on the future of geometry processing.
2
4
62
My best predictor of ICML submission success has been whether the algorithm has a cute acronym.
3
2
61
Just reserved the @MITMuseum and @bostonteaship for the Symposium on Geometry Processing (SGP) 2024 social events! Get excited: (registration opens soon!). And, please reach out if your company or organization is interested in sponsoring!.
0
9
60
Fantastic work driven by our brilliant student @ana_dodik — tackling classical geometry processing problems using neural representations! Check out her paper at @SIGGRAPHAsia 🦘.
Can neural fields help unlock new understandings of an old geometry problem? Excited to announce our latest SIGGRAPH Asia work: Variational Barycentric Coordinates! 🧵. - with @OdedStein @vincesitzmann @JustinMSolomon
0
4
59
The number of colleagues, students, & others who volunteered time/resources to support SGI is inspirational and reflects strongly on the geometry processing community. Here’s a list: — SGI’s success is completely due to their effort & enthusiasm! ❤️.
0
6
60
More late-breaking posts on the SGI 2022 blog! Geodesic distance approximation, neural implicit boundary representations, mesh denoising, and more. 🧑🏫👩🔬🧑💻 So proud of our SGI Fellows' work this year! NB: Many are on the market for a PhD position. .
0
8
59
Academic life advice needed! This year, was so focused on surviving the tenure process, I didn’t realize I can take a sabbatical shortly thereafter. Turns out I can take a year of sabbatical starting fall 2024 (exciting!), but I have no idea how to plan. Questions ⬇️.
5
2
57
Loved seeing everyone at SIGGRAPH! Many thanks to those who attended my award talk on Wednesday---meant a lot to have you there. There might be a video eventually (I'll be too self conscious to watch!), but for now here's the text: ---scripted this time. .
1
1
58
A lot of MIT press for our group lately! This time our student Yue’s work on deep learning from point clouds: — Yue’s been our long standing expert in 3D/point cloud deep learning, and his ideas have already led to 3 great publications in this area! 👨💻.
1
14
55
This chair is lecturing to an audience of 100+! Teaching intro to computer graphics with pre-recorded videos, plus live Q&A in a Slack during viewing parties and after-playback Zoom discussion time.
5
0
57
Obscure writing rules I feel strongly about for no rational reason:.• utilize→use.• in order to→to.• comma before “which” but not “that”.• comma before “and” when connecting independent clauses but but not dependent clauses.• Oxford comma=💕.None of these actually matters.
5
0
54
Puzzled that a colleague at another institution would waste their time trolling the comments in the lecture 19 video of a niche geometry course, posted for free on YouTube.
2
0
56
🚨Attention faculty, researchers, and grad students in geometry and adjacent fields (broadly considered): The Summer Geometry Initiative needs your help!🚨
3
14
55
Pressed for time but want to learn a graceful bridge between vector geometry and raster data in deep learning? Here's a 1min overview of @MIT_CSAIL student Dima Smirnov's CVPR paper "Deep Parametric Shape Predictions using Distance Fields" w/@AdobeResearch
1
8
56
Hot take: Journal/conference review forms with a million short answer questions are detrimental to the review process, because they encourage lots of short+thoughtless+superficial responses instead of a detailed review. Plus they don't help the editors on the receiving end.
2
3
56
The SIGGRAPH talk was like a historical retrospective. 🤓📚 In addition to his brilliant research, @yuewang314 is one of the most motivated students I’ve worked with, and he didn’t let rejection get him down—polished DGCNN while doing more projects expanding his research program.
The "Dynamic Graph CNNs" paper was rejected from SIGGRAPH '18 and presented at the same venue just 2 years later! Also, it currently has >700 citations. "Don't let a rejection determine the worth of your paper" - I learn this everyday in different ways. :).
1
9
56
Has ghosting become a common, socially accepted phenomenon post-Covid? i.e., making/confirming plans and then vanishing at the last minute? Have noticed this professionally and in dating, curious for an explanation (or maybe it’s me). Sincerely, this grumpy/frustrated old man.
10
4
55
Pet peeve: Papers whose list of contributions includes "We are the first to [do niche task] using [specialized tool/method]." Being first is rarely a contribution in itself---communicate the interest rather than making a territorial claim. (Some of my papers violate this. Oops.).
5
1
52
Hi friends! At this point, I've personally contacted 70+ folks and am having a tough time identifying sponsors for these programs (tough year in industry). I need your help! If you have contacts in relevant companies or foundations, please make an intro here or via email. Thanks!.
2024 will be a big year for geometry at @MIT_CSAIL/@MITEECS: We'll be organizing/hosting BOTH the Symposium on Geometry Processing (SGP) and Summer Geometry Initiative (SGI)!. We are excited to bring geometry to @MIT, but it will come at substantial cost---so we need your help!
0
20
51
I ask all our students to do this, and they often end up in mildly painful meetings where I’m reading their papers out loud. Plus there’s nothing more enthralling than a dramatic reading of 12 pages of measure theory. Best way to catch typos and smooth out awkward writing.
Free writing tip: read your papers *out loud* to find the bad and awkward-sounding sentences. (This may be most effective for native English speakers. Would be curious to hear if others find it useful.).
3
1
56
Last day at the Oberwolfach Research Institute for Mathematics in 🇩🇪—one of my favorite places on earth. Beautiful institute, chance to catch up with collaborators, unlimited coffee, and even a recital hall with a cello to play! Heading home after a quick stop in Zurich tonight.
2
0
52
🚨🚨 We are seeking a postdoctoral associate starting fall 2023 to carry out research in a collaboration between the @MIT_CSAIL Geometric Data Processing group and the @MITIBMLab. 👀💻.
1
14
51
Is your reading group/seminar interested in reviewing numerical methods (matrix factorization, iterative methods, PDE/ODE, quadrature) from an applied perspective? I could use some folks to beta test the 2nd ed of my "Numerical Algorithms" textbook. Ping me if you're interested!.
3
4
49
My feelings about in-person teaching starting in an hour.
0
0
51
Excited to host @mmbronstein in our graduate 6.838 Shape Analysis course today! Not the usual distinguished classroom setting, but it’s great he can “drop by” without crossing the Atlantic to share his expertise on geometric deep learning.
0
0
52
Applied optimal transport — 72 packages ready to mail to our SGI Fellows, volunteers, and mentors!
2
0
52
Positions confirmed! Still hoping to hear from more great candidates at the intersection of ML and geometric data. Please send me a quick email if you’re interested or have questions! And, share with any interested colleagues or students.
Our group in @MIT_CSAIL will likely (finalizing funding) have an opening for a postdoc at the intersection of ML and geometric representations (eg meshes, CAD, LiDAR) to start late 2020/early 2021. Check out our web page to get an idea of what we’re up to:
0
14
50
Does your company have some extra funds for outreach left to spend before the end of 2023? If so, I'd suggest sponsoring the Summer Geometry Initiative ( or Symposium on Geometry Processing ( 2024. Contact me for more information!.
0
14
48
In what can only be described as the “best day ever” in the @MIT_CSAIL Geometric Data Processing group, an anonymous “Panini Pal” saw this post and mailed my grad students a panini press! Thanks, mysterious Panini Pal!
2
2
48
New SGI mug just dropped! Literally, thanks to a broken cart wheel. Volunteer or visit us at MIT to collect them all! (While supplies last.)
3
4
49
I’m always amused to read “traditional” in papers referring to algorithms from one year ago, as if they were designed by candlelight in a simpler age. (Suggested rewording: “conventional”).
1
1
49
So proud of our SGI students’ progress! Week 4 of 6 and they’re implementing new methods for fluid sim. Check out @natashadiederen, Joana, Lucas, and @paulkry’s blog post here: (and take a look at posts reporting on other projects while you’re there!).
0
3
48