| 1 |
* |
| 2 |
^ Ideas |
| 3 |
a) Learning for AI Photos Geolocations. |
| 4 |
b) Find sth according to a photo or video. |
| 5 |
- google and yandex photo work badly. |
| 6 |
c) AI for cybersport. |
| 7 |
d) Mapping for local places of cities |
| 8 |
e) Including some control systems in different medical organizations (doubtful)9 |
| 9 |
f) Recording of Podcasts |
| 10 |
g) making flyers, logos, banners via AI |
| 11 |
h) making skins for games |
| 12 |
- need substans painter and 3d |
| 13 |
i) making sites vie AI |
| 14 |
k) making educational courses |
| 15 |
* |
| 16 |
^ Courses AI |
| 17 |
a) Deeplearning.ai/courses --https://www.deeplearning.ai/courses/machine-learning-specialization/ |
| 18 |
- Strong |
| 19 |
** |
| 20 |
* General |
| 21 |
^ |
| 22 |
a) Palantir works with huge, massive of data from different sources. |
| 23 |
- it allows users to find connection between objects |
| 24 |
b) Usual sources are texts, audio and videos. |
| 25 |
c) No need programmers and engineers because all process in intuitive graphic interface. |
| 26 |
- requires form via natural languages. |
| 27 |
^ Gotham -- a project for government special services |
| 28 |
a) initially is called Government |
| 29 |
b) Gotham uses technique of ontology |
| 30 |
c) Can have methods of semantic search |
| 31 |
- can also have phonetic search |
| 32 |
d) Uses genetic algorithm |
| 33 |
e) it has means of loading data, inserted in service FEI (front-end import) where it can semiautomatically take information from web-sites |
| 34 |
^ Ontology in products of Palantir can by one of three types of data |
| 35 |
a) entity -- subjects or objects of real world |
| 36 |
b) events -- actions on entities at certain period of time or at certain point of space |
| 37 |
c) documents -- confirming information about real world, combined in determined format (uses HTML). |
| 38 |
^ Additions |
| 39 |
a) it might use GPT-4 |
| 40 |
b) better make requires with inserting limited text |
| 41 |
* |
| 42 |
^ Questions |
| 43 |
a) LLM -- what's that |
| 44 |
- https://www.youtube.com/watch?v=2p0WcDp9LBc |
| 45 |
' on the 2nd min. |
| 46 |
*** |
| 47 |
For my future PC |
| 48 |
- RTX 4090 Supreme: 230 000 |
| 49 |
- Ryzen 7950x: 70 000 |
| 50 |
- 2000 ГБ SSD M.2 накопитель Gigabyte AORUS Gen5 10000 [AG510K2TB] -- 35 000 (2 Штуки надо) |
| 51 |
- 2 монитора на 17 дюймов фул нд ~ 25 000 (2 штуки надо) |
Комментарии