Update summaries-furtado.md

_objects/summaries-furtado.md

@@ -66,3 +66,18 @@ The lesson itself revolved around the topic: Why are networks important and why
 We collected ideas on the topic of wedding planning. Through the simulation (with the table extension) that we created, we see that there are already existing relationships, but there are also people who do not know each other yet. We can group them by age, languages, interests and relationship status, to see which people can be divided into which table. Left is always a group of "odd one" (which is not easy to put in a group). There are different dynamics in different tables. In the sense of relationship, communication and dominance. Someone can control multiple conversations; who talks to whom and who limits themselves to a few people. When different groups meet, different information is exchanged, but this does not necessarily mean that this information reaches everyone in the group. In networks there are also brokers. They are powerful (they have control, can decide whether to pass on information or keep it to themselves.) Over time, structures and dynamics can change. New network can emerge, e.g. some find themselves in the wedding. (Lovers) 
 Afterwards we had a PowerPoint presentation. Where we learned what a social network. A specific set of links. (we saw different types in class) We can use it to explain why people behave differently in different groups. (Ex: smoking behavior, voting behavior etc.) Our behavior can also destroy networks. Relationships are important and how you are connected to them. Networks are flexible. Cases can be represented visually and networks are not limited to people. Social network analysis is both theory and method. Today, everything can be defined as a network. (expl: Family network) From the historical perspective, network can guide us to the self-perception of historical actors, to the metaphor for a group, to the constructed scientific research object and to the theory/method driven research approach.
 Finally, we had the task of creating our own wedding list. My group created this on Excel and used the table in Palladio. Thus, we created our own fictional network.
+
+
+---
+title: Session Summaries by Emilie-Furtado
+abstract: Summary-6
+authors:
+  - emifurtado
+date: 2024-10-30
+---
+## EU Archive, 30-10-2024
+For this class, we received a presentation from Ludovic Delepine and Marco Amabilino from the European Parliament Archives. First, we were introduced to the team and then we watched a video that we had to see before the lesson. In the presentation, they explained the steps they had to take in developing the website. First, they wanted to make the documents managed by the Parliament more visual and accessible. Previously, the only way to access the documents was by email or in person. Technically speaking, they are a small team. They wanted to solve this problem digitally, so they needed a server and at the same time, the documents should be accessible from anywhere from now on. That's how Archibot was born. Important figures were introduced such as Edgar F. Codd. In the beginning, documents were just pics because they were scanned. Now they need a solution to make the characters readable/recognizable. Another important figure who was introduced is Gerard Salton.
+Next, we learn that metadata is generated = deep learning. A model is trained to generate metadata.
+Next, they developed their own "ChatGPT" = "Ask the EP Archive". The idea was to create a simple search file where you can ask questions and get an answer. It works in different languages and you get the answer in the language you ask in, even if the document is not available in the same language. ChatGPT is unreliable and sometimes gives made-up answers. They create a prompt that says it is unable to answer a question. They have also added to the system that it should not engage in discussions. The data is not shared, everything stays in their system and runs there. The system selects the 10 most important documents. If you want all the documents on a topic, you have to search yourself. We were also advised not to enter any personal data into the system.
+Questions were asked and answered in the discussion round. For example, we learned that translation is based on training. Some languages are better trained than others, so if they can't translate something, it is automatically translated into English. We also learned that they use Amazon because of the contracts. We also talked about the main fears and prejudices. People are afraid of being replaced by machines. Humans and AI make mistakes. Both can support each other. Humans can spread information, just like AI (which can also spread wrong information), even professors have wrong answers. Nobody is perfect, even AI is not perfect. AI is not meant to be perfect or replace humans. They are a tool to promote and facilitate searching. They don't want to replace the archive. 
+