![[CESERH] IMACTIS Research Seminar. Images Today org. by DONDERO & D’ARMENIO (21/9/21) - Seance 3](https://i.ytimg.com/vi/1-rQLLchBdA/maxresdefault.jpg)
[CESERH] IMACTIS Research Seminar. Images Today org. by DONDERO & D’ARMENIO (21/9/21) - Seance 3
[CESERH] IMACTIS Research Seminar. Images Today org. by DONDERO & D’ARMENIO (21/9/21) - Session 3
Keywords
Summary
160 words
Critical Evaluation
Value of the Information & Strength of the Argument
The talk offers significant value by documenting the real-world difficulties and solutions in a large interdisciplinary research project. Leone argues convincingly for the integration of AI into humanities research, addressing both legitimate concerns (biases, limits) and outdated resistance. He provides concrete examples of obstacles (legal, infrastructural) and how they were overcome through collaboration. The argumentation is coherent and personally grounded, though it leans on anecdotal evidence rather than systematic data. The discussion of dataset creation and GDPR compliance is particularly insightful, highlighting often-overlooked ethical and practical aspects.
Scientific Rigor, Source Quality, Title Accuracy
The presentation is academically rigorous, based on the speaker’s extensive experience and project documentation. However, no external sources are explicitly cited in the video; the content relies on project details and personal reflections. The title is somewhat cryptic but aligns with the seminar’s theme of images and digital studies. The audience is likely specialized, as indicated by the technical terms and academic context. No comments were provided, so audience reception cannot be assessed.
175 words
Title / Content Match
The title accurately reflects the seminar context and the speaker's focus on images and digital semiotics, though the acronyms and session number make it less descriptive than the actual content.
Quality & Reliability
7/10
Presentation by a full professor of philosophy of communication and semiotics, discussing his ERC-funded project FACETS. The content is based on his own experience and project details, credible but not peer-reviewed. Some claims about AI limitations are general and reflect personal perspective.
Key Moments
Markers derived by PSI from the transcript: the creator did not define chapters.
- Introduction of Massimo Leone and his ERC project FACETS
- Discussion of initial challenges with big data and Amazon Mechanical Turk
- Collaboration with Polytechnic University of Turin and creation of FRESCO
- Explanation of datasets, GDPR constraints, and profile picture analysis
- Reflections on merging semiotics with AI and overcoming academic conservatism
Contribution & Novelties
The talk’s originality lies in presenting a pioneering attempt to apply AI to visual semiotics, specifically for large-scale analysis of facial images. It offers a candid account of the challenges, including ethical and legal barriers, and demonstrates how collaboration with engineers reshaped a humanities project. The speaker’s emphasis on teaching AI to ‘see’ semiotically is a novel concept.
Pour aller plus loin :
- Semiotic square — Conceptual foundation for the ‘digital square’ mentioned.
- FACETS project on ERC — Official EU project description, though exact page may vary.
- Generative adversarial network — AI technique used to generate synthetic facial images.
- GDPR — Key legal framework affecting the handling of personal image data.
111 words
Radar Profile
The profile shows relatively high scores across all dimensions, with the highest in information quantity and quality, reflecting the depth and academic value. Slightly lower scores in technical level and reliability suggest the content is accessible but not deeply technical, and relies on the speaker's personal account rather than peer-reviewed evidence.