Having done prior research on symbolic programming for agents, qualitative simulation (QSIM, QPM), and ontologies, I wanted to integrate CLIPS with the tremendous L4T CV and D/CDNN development environment.
This was little problem with my Xeon 20 + RTX NVIDIA SDK workstation using standard Ubuntu 18.04 software sources, but there were no ready binaries availability for the Xaviers.
It was back to traditional compile from the latest 2021 C source
[CLIPS Rule Based Programming Language - Browse /CLIPS/6.40 at SourceForge.net]
using the NX (ARM64) gcc. The provided Linux makefile had to be modified for the appropriate OS version.
It runs perfectly through the familiar terminal interpreter and the CLIPS C library was created. Work is proceeding on similar ARM64 specific clipspy and CLIPSJNI (Python and Java bindings) like those also installed on the Xeon workstation.
Speaking of JDK, having a good OpenJDK and OpenJRE installation for L4T made it possible to run Stanford’s Protege to edit and reason with OWL ontologies. Minor adjustments were needed for launching the generic Linux .jar.
Not incidentally, the NIH’s widely used ImageJ Java app can read and process just about any image array, including BIOFORMATS
Why all this? How about for using the medical standard SNOWMED-CT ontology for semantically linking DNN-segmented structures for symbolic reasoning about brain tumors in MRI image stacks…
Hi @trelease, thanks for sharing - sounds like an interesting project! Glad that you were able to get it compiled and running. Let us know how it goes!
Thanks for the response. And thanks for all the great Jetson ML/CV developer education resources you’ve provided as Jetson edu lead over these years!
After 4 decades working in previous ‘waves’ of AI, it is amazing to be able to develop and to experiment with such powerful real working connectionist hardware on the edge. The GTC sessions on medical imaging and healthcare have been outstanding, showcasing the tremendous advances in GPU-based DL and scientific computing!
And after working a couple of years on biomedical imaging with my AGX Xavier dev kits, I am just boggled by the specs for the upcoming Orin!
Hope to have more to share on the Forums as all the years’ pandemic isolation work continues to come together…
Department of Pathology and Laboratory Medicine and
UCLA Brain Research Institute
David Geffen School of Medicine at UCLA
PS Since my lab-computing focused work has a different background than others, I’m including a few prior pubs for off-forum context.
TreleaseEtAl1999.pdf (1000 KB)
Trelease&Rosset2008.pdf (391 KB)
Trelease2006.pdf (799 KB)
Thanks @trelease, it’s our pleasure to help with the tutorials and educational content to support other developers like yourself. Very interesting research you are doing - and for a great cause. Keep up the great work!
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.