Universität Wien

160167 UE Web APIs and large language models in humanities research (2023W)

5.00 ECTS (2.00 SWS), SPL 16 - Musikwissenschaft und Sprachwissenschaft
Continuous assessment of course work
Moodle

Registration/Deregistration

Note: The time of your registration within the registration period has no effect on the allocation of places (no first come, first served).

Details

max. 30 participants
Language: English

Lecturers

Classes (iCal) - next class is marked with N

  • Thursday 12.10. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 19.10. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 09.11. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 16.11. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 23.11. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 30.11. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 07.12. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 14.12. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 11.01. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 18.01. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG
  • Thursday 25.01. 18:30 - 20:00 Hörsaal 1 Sensengasse 3a 1.OG

Information

Aims, contents and method of the course

The rise of cloud technologies, particularly Large Language Models (LLM) like ChatGPT, heralds a promising new frontier in digital humanities research. Rather than merely focusing on technological innovations that offer fresh insights into data, we now have the means to automate intricate, in-depth analysis that resonates with the primary interests of humanities scholars.

In this course, participants will learn to develop sophisticated Web API-based applications. These applications will harness data from sources such as Europeana and Wikidata and utilize versatile data-processing tools like ChatGPT to deliver profound insights from extensive datasets.

The following prerequisites and guidelines are imperative:

*Nosce Pythonem:* Proficiency in Python is crucial.
*Bring your own computer:* Participants must bring their personal computing devices to the sessions and should be proficient in installing PyPI packages via pip or conda on these devices.
*Bring your own data:* Participants should select a data source, either of academic or personal relevance, that they want to explore using Web API and LLM technologies. As a guideline, the data should be in a machine-readable format and exceed 150,000 characters (around 2x Le Petit Prince).
*Procure your API key:* It is the responsibility of students to register for their API access and bear any associated costs. The services will primarily involve OpenAI, but also other APIs tailored to their specific project requirements. The total expenditure throughout the term is typically under 20€.

Assessment and permitted materials

- Programming exercises 60%
- Project 40%

Minimum requirements and assessment criteria

>= 90% very good (1)
>= 80% good (2)
>= 65% satisfactory (3)
>= 50% sufficient (4)
< 50% not sufficient (5)

Examination topics

Students must:
- Demonstrate proficiency in the technologies via the coding assignments, and
- Complete a project which conducts non-trivial processing on their chosen data source in a "proof-of-concept" manner. This means that while the essential steps should be present and the workflow should yield the intended type of outcome, the results are not required to be flawless or ready for practical application.

Reading list

Students are encouraged to self-study through online resources, particularly those related to LLMs, prompt engineering, and notably, LangChain.

Association in the course directory

MA DH: DH Skills II; S-DH Cluster 1
BA Sprachwissenschaft: Alternative Erweiterung

Last modified: Th 21.09.2023 17:27