Dataview

LIST
FROM #Collection
WHERE file.name = this.Entry-For

Insert entry name


🎤 Vocab

❗ Information

Small summary

✒️ -> Scratch Notes

Manos Kukomudas
OUMI - open universal machine intelligence

change in AI:

  • open -> closed
  • controlled by elite tech gians
    why?
  • access to compue
  • platform to make AI resaerch easy, efficient, cumulative
  • “hard fo rstudents t do this research. none has managed to do distributed multi-mode training”
  • mindset: people feel like they cant compete

Deepseek wakeup call

  • US investing far more on closed source than open models
    Truly open - open data, code, models, and collaboration
    similar to Linux >> Unix
  • better with time
  • flexibility / support valued

People have cam etotgether before

  • Humand Genome Project / CERN LHC
  • why not for AI?

OUMI Resaerch Platofrm

  • provide platform for AI
  • Foundation models research
    • Building models is complex:
      • Pretraining, supervised fine tuning, preference tuning, evaluation, data synthesis, inference, deployment
      • Models, datasets, optimization techniques, benchmarks
    • fully open ource (apache 2.0)
    • built on pytorch, optimized

For this hackathan:

  • easy to experiment any large model in oumi
  • ustomizea smal model with ur own data that fits on rasberry pi
  • or train larrge model and distill to a small model that fits

Judging Rubric

4 pts - define problm
4 pts - integrate hardware well
4 pts - leverage data / computation

  • potentially ucd ag ml
  • potentially make your own
  • demonstrate this all in presentation
    • reputable?
      2 pts - presentation quality

Demo

Zhaodan and Prabash
2 kits:

  • Sprysense
    • low power
    • main board & extension board
      • GPS/SSD/Audio
      • Sim cards/Wifi
      • Agent pulse? make and deploy lightweight ML models on edge devices
  • Rassbery pi AI Camera
    • more powerful, made for small NN
    • all processing on camera iteself
      • rpi0 free for other processing, not just model
    • rpi0
      • wifi
        use at least 1

team 5 us

🧪 -> Refresh the Info

Did you generally find the overall content understandable or compelling or relevant or not, and why, or which aspects of the reading were most novel or challenging for you and which aspects were most familiar or straightforward?)

Did a specific aspect of the reading raise questions for you or relate to other ideas and findings you’ve encountered, or are there other related issues you wish had been covered?)

Resources

  • Put useful links here

Connections

  • Link all related words