Auto saved by Logseq

2025-06-02 17:15:13 +02:00
commit d9ea3d552d
6279 changed files with 711145 additions and 0 deletions
@@ -0,0 +1,59 @@
+-
+- ## Links
+	- [Program - MODELS 2023 (researchr.org)](https://conf.researchr.org/program/models-2023/program-models-2023/?prog=Your%20Program&past=Show%20upcoming%20events%20only)
+	- Proceedings: [MODELS 2023 (computer.org)](https://conferences.computer.org/modelspub/#!/home)
+- ## General notes / [[Ideas]]
+	- #promptengineering
+	- Composition of different recommender systems (look at what Juan is doing)
+		- ![sle23-paper78.pdf](../assets/sle23-paper78_1696238141035_0.pdf)
+		  collapsed:: true
+			-
+	- Can we consider LLMs as generic execution environments that can take two different inputs:
+		- Examples or specifications of the goal-specific execution environments
+		- Take input consistent with the previous specification and thus able to produce domain / goal specific outputs
+	- ## **Best LLM APIs on the market**
+		- While comparing LLM APIs, it is crucial to consider different aspects, among others, cost security and privacy. LLM experts at Eden AI tested, compared, and used many LLM APIs of the market. Here are some actors that perform well (in alphabetical order):
+			- AI21Labs: Jurassic-2
+			- Anthropic: Claude 2
+			- Cohere
+			- Falcon
+			- Google Bard: LaMDA
+			- Google Generative: PalM
+			- Meta: Llama2
+			- OpenAI: GPT
+		- [Best Large Language Model APIs in 2023 | Eden AI](https://www.edenai.co/post/best-large-language-model-apis)
+	-
+- ## TODOs
+	- {{query (and (task TODO) [[CONFERENCES/MODELS2023]])}}
+- [[02-10-2023]]
+	- ![image.png](../assets/image_1696236044974_0.png){:height 274, :width 756}
+		- ![image.png](../assets/image_1703840397306_0.png){:height 766, :width 666}
+	- [Encoding Conceptual Models for Machine Learning: A Systematic Review562](https://conferences.computer.org/modelspub/pdfs/MODELS-C2023-6qG8PWpWRRmVmqYwIhdqkX/249800a562/249800a562.pdf)* Syed Juned Ali (TU Wien), Aleksandar Gavric (TU Wien), Henderik Proper (TU Wien), Dominik Bork (TU Wien)*
+		- [[@Encoding Conceptual Models for Machine Learning: A Systematic Review]]
+	- [[@Extracting Domain Models from Textual Requirements in the Era of Large Language Models]]
+	  collapsed:: true
+		- DONE Look at function_calls mechanism introduced by ChatGPT late June 2023
+		- ![IMG_20231002_112815.jpg](../assets/IMG_20231002_112815_1696239236860_0.jpg)
+			- DONE Look at what CRF is
+		- ![IMG_20231002_113006.jpg](../assets/IMG_20231002_113006_1696239225018_0.jpg)
+	- [[@Prompting or Fine-tuning? A Comparative Study of Large Language Models for Taxonomy Construction]]
+	- ### Lightining talks
+		- ![IMG_20231002_154636.jpg](../assets/IMG_20231002_154636_1696254634766_0.jpg){:height 541, :width 886}
+		- ![IMG_20231002_154434.jpg](../assets/IMG_20231002_154434_1696254650629_0.jpg){:height 555, :width 887}
+		- ![IMG_20231002_155856.jpg](../assets/IMG_20231002_155856_1696255256020_0.jpg){:height 477, :width 890}
+			-
+				-
+		-
+- [[05-10-2023]]
+	- ### MODELS2023 PANEL
+	  id:: 651a8a88-9694-4887-9d92-fa2364e9b940
+	  collapsed:: true
+		- **Democratization of LLMs-based tools** #challenges #LLMs
+			- Lock-in problems
+			- Costs
+			- Risk of increasing the gap among people depending on their economical capacities
+		- **Is AI reserved only for big players?**
+			- Houari answers by saying that big players are developing "Assembly" for us
+			- Daniel answers that they are not caring about security, safety, etc. of our society. We do.
+		- **Do we need models considering what AI will be in 5 or 10 years?**
+-