Choose Your Weapon: Survival Strategies for Depressed AI Academics

 https://arxiv.org/abs/2304.06035

There are a lot of suggestions in this paper that I think are helpful In particular,

  • “In-the-wild domains such as games, IoT, and autonomous vehicles could allow AI to be deployed next to their end user and the data the user generates, i.e. at the edge of the network. This is often called edge AI [8], the operation of AI applications in devices of the physical world is possible when memory requirements are low and inference occurs rapidly.” (Togelius and Yannakakis, 2023, p. 4)
  • “Solve Problems Few Care About (For Now!)” (Togelius and Yannakakis, 2023, p. 5)
  • “In academia, failure can be as instructive and valuable as success and the stakes are overall lower. Many important inventions and ideas in AI come from trying the “wrong” thing. In particular, all of deep learning stems from researchers stubbornly working on neural networks even though there were good theoretical reasons why they shouldn’t work.” (Togelius and Yannakakis, 2023, p. 5) 
  • “At the most basic level, open-sourcing models, including both weights and training scripts, helps a lot. It allows academic AI researchers to study the trained models, fine-tune them, and build systems around them. It still leaves academic researchers uncompetitive when it comes to training new models, but it is a start. To their credit, several large industrial research organizations regularly release their most capable models publicly. Others don’t, and could rightly be shamed for not doing so.” (Togelius and Yannakakis, 2023, p. 7) 

 

Comments

Post a Comment

Popular posts from this blog

Communicating and downloading data from the James Webb Space Telescope

We use the DSN ([Deep Space Network](https://www.jpl.nasa.gov/missions/dsn)) to communicate with the observatory. We receive data when we have a contact with Webb using a DSN antenna The DSN has three sites around the world, each positioned 120 degrees apart. There are antennas in Goldstone, California; Canberra, Australia; and Madrid, Spain. This allows us to communicate with Webb at any time of day Each of the [DSN complexes](https://www.nasa.gov/directorates/heo/scan/services/networks/deep_space_network/complexes) has different types of antennas, including 70-meter (230-foot in diameter), 34-meter (111-foot in diameter), and 26-meter (85-foot in diameter) antennas. The DSN complexes use the 34-meter antennas to talk with Webb with the 70-meter antennas as a backup. The DSN supports different radio frequency allocations, such as the S-band and Ka-band frequencies that Webb uses. -band has a lower bandwidth, and we use that to send commands to the spacecraft (e.g., start recorder play...
Read more

Zotero to Tiddlywiki5

  Zotero is a great tool but it has been challenging to look up bib and notes taken in Zotero on Tiddlywiki. It is really inconvinient to have to always have to switch between these two apps. Therefore, I wrote this python script which takes the csv that you get from running Zotero's default csv translators and converts it into something that you can import to Tiddlywiki directly. Note that Tiddlywiki does not allow you to import arbitrary JSON files such as   {     'x': 10,     'y': 20 } as new tiddlers. Instead any JSON file that you import must have the following keys to be loaded as a new tiddler {     'text'        : '' ,     'tags'        : '' ,     'modified'    : '' ,     'created'     : '' ,     'title'       : '' } ...
Read more