HeyJaredAI in Engineering, Physics, Aerodynamics
Read what I read, code what I code. I share lots of papers/code for topics in artificial intelligence/machine learning applied to engineering (CFD, FEA, numerical methods, aerodynamics, turbulence modeling, & more). I will publish enough to keep you busy.
- Indexed issues, last 90 days
- 12
- Latest publication
- Sep 29, 2026
- Audience
- Checking…
- Earliest in this view
- Aug 31, 2026
Latest issues
Robotics News & Literature (September 29th) (opens the original)
Read excerpt
A robot can finish a task once and still be difficult to improve. Its successful demonstration may cover only one arrangement of objects, its failed attempts may be discarded, and its controller may struggle with disturbances absent from training. This week’s research asks how to make that experience more useful, and how to tell whether the resulting improvement survives deployment.The five developments below examine training data, feedback control, and evaluation. Part 2 pairs each with two pap
Bumper Beam Crash Dataset (& Training a Transformer on it) (opens the original)
Read excerpt
<img alt="" class="sizing-normal" height="630" src="https://substackcdn.com/image/fetch/$s_!GJsR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd5ac0205-345d-41a6-ad36-dec4aa07bfa2_1120x630.gif"
Machine Learning Surrogates for Engineering (opens the original)
Read excerpt
I have been writing a lot of reviews and thought it was high time to make good shortlist of some of my favorite models. Following that, i’ll provide another literature review for ML surrogates over the past week (including both academic and industrial news).Advanced CFD and FEA surrogate highlightsFor my shortlist, these four papers from the past year deserve attention. They address industrial geometry and simulation fidelity, but their strengths differ. Dates below are first arXiv releases; the
Robotics News & Literature (September 21st) (opens the original)
Read excerpt
A robot can recognize a plug and still fail to insert it. It can learn to fold a towel in one room and struggle in the next, or lose track of an object the moment it disappears into a box. This week’s research gets specific about these gaps: what training experience transfers, what information a controller needs, and how quickly that information must reach the motors.The five developments below are useful because their experiments expose a design choice. Each is followed, in Part 2, by two paper
Physics AI/ML Literature Review (through September 18th) (opens the original)
Read excerpt
Company updates first, then papers with PDFs. Techie pick before we start - a bit ago I announced a competition that launched for who could build the best ML surrogate on this aerospace problem. Results are in, competition is over, and we can review who won! <img alt="" cla
Publishing over time
Last 90 days. Choose a month to open its work.
Recurring subjects
Named in the text we hold. One piece can cover several.
Audience
No verified audience measurement yet.
About this data
Counts cover the work we have indexed. Tone needs enough text and a confident classification. Excerpts and episode notes are not full articles or transcripts.
Identity or attribution wrong? Suggest a correction.