Undergraduate Seminar in Data Science

Monthly undergraduate data science seminar with featured talks, video archive, and recent channel uploads.

Intro

Monthly seminar with innovative ideas and trends in Data Science applied to industry for undergraduate students.

Official channel:

Highlights

Featured seminar session

Featured seminar visual

Session title

Deep Mutual Information Meets Segmentation by Felipe Felix

This talk introduces a method that improves neural-network segmentation by focusing on the learning objective instead of changing the architecture. Using mutual information as a learning goal, an auxiliary network enhances the information shared between input and output, refining weight updates. The result is better capture of relevant relationships and improved segmentation performance across different object sizes.

  • Platform: Microsoft Teams
  • Access: Session links are shared through official seminar announcements.
  • Related publication: [10.1007/978-3-032-08704-1_22](https://doi.org/10.1007/978-3-032-08704-1_22)
  • Related project area: [Research Projects](/pages/research/projects/)

For upcoming sessions and updates, follow the channel and seminar announcements.

Seminar video archive

Use the embedded playlist below to browse previously published videos directly from the website.

Recent uploads

Latest videos from the channel

Seminar objectives

The seminar is designed to expose undergraduate students to current topics in data science, including:

  • machine learning foundations
  • data ethics and responsible AI
  • mathematical and statistical methods
  • real-world applications and case studies

Call to action

If you want to participate as a speaker, collaborator, or student, please use the [Contact](/pages/contact/) page.

References

  • Official channel: [@UndergraduateDataScience](https://www.youtube.com/@UndergraduateDataScience)
  • Main uploads playlist: embedded above in this page
  • Program context and related outreach: [Blog](/pages/social/)

Related content

Site sections

Profiles