Konstantin Kobs

Data Scientist

[Konstantin Kobs @konstantinkobs]

@datascience_jmu Congrats! 🎉

[jan @janpf95]

All good things come in threes: We will present our last paper at #NAACL2024 "The Roman Empire Strikes Back" at the @SemEvalWorkshop poster session at 14:00. Meet us there!
It is a joint work with @konstantinkobs on detecting hallucination detection in text generation models.

[Jasper Polak @polak_jasper]

Dunno who made this but it's genius. Why change fails:

[Happy PhD Supervisor @HaPhDsupervisor]

A PhD student who recently finished her dissertation gave me the book “Big Panda and Tiny Dragon” from James Norbury. This page resonated so much. Does it also apply to your academic journey or would you disagree?

#AcademicTwitter #PhDlife #PhDchat

[BibSonomy @BibSonomyCrew]

We're pleased to announce the release of our #ChatGPT Plugin🎉
Designed for researchers familiar with https://t.co/iwk6dbUBeZ and those looking to start, this tool grants access to a database of 200M+ scientific publications! (🧵👇🏻1/4)

[Konstantin Kobs @konstantinkobs]

@Sagar_Vaze @SergeBelongie @KateSaenko @drfeifei @LiJunnan0409 @kohjingyu @rsalakhu Nice work! We introduced a very similar setting with the same motivation in this year's WACV paper "InDiReCT: Language-Guided Zero-Shot Deep Metric Learning for Images". Might be of interest to you for future work on this topic, since we propose a different approach. 😊

[Konstantin Kobs @konstantinkobs]

@opensourcesblog @OpenAI I think this will get solved when it is hooked up to a calculator and a web browser. I think it will be much more powerful then.

[Konstantin Kobs @konstantinkobs]

Our paper “CLIP knows Image Aesthetics” is now published. We show that CLIP is a better base for Image Aesthetic Assessment models than ImageNet, providing useful results even in a zero-shot setting. Previous CLIP studies focus on content and not style features. @datascience_jmu

[Frontiers in AI & Big Data @FrontAIBigData]

New Research: CLIP knows image aesthetics https://t.co/Lmtd1lD8h1 #AI

[Konstantin Kobs @konstantinkobs]

@MushtaqBilalPhD We also developed https://t.co/mfMEOfb7Om to recommend ML conferences and journals based on title and abstract. It also highlights words and phrases that were important for the recommendation.

[Data Science - Professor X @datascience_jmu]

Our paper "InDiReCT: Language-Guided Zero-Shot Deep Metric Learning for Images" has been accepted at WACV 2023 https://t.co/M66U5AXZ47

In this paper by K. Kobs, M. Steininger, and A. Hotho, we use language to guide an image embedding process such that the resulting embedding s…

[Data Science - Professor X @datascience_jmu]

Our paper "On Background Bias in Deep Metric Learning" has been accepted to ICMV 2022 https://t.co/X6syUxk5bG

In this paper, we investigate if Deep Metric Learning models are prone to background bias and test a method to alleviate such bias.

[Konstantin Kobs @konstantinkobs]

@DanielZuegner @guennemann Congratulations! 🎉

[Konstantin Kobs @konstantinkobs]

@mlopezantequera @ICCV_2021 @datascience_jmu We found a significant difference between classification and ranking based loss functions, both on pixel-level as well as image property level. 😉

[Konstantin Kobs @konstantinkobs]

I am so excited to announce that our paper "Do Different Deep Metric Learning Losses Lead to Similar Learned Features?" got accepted at @ICCV_2021! 😍 @datascience_jmu

[Konstantin Kobs @konstantinkobs]

@routenote First song using RouteNote as distributor! 🎉 https://t.co/Xw8oyVbiiV

[Khoa Vu @KhoaVuUmn]

PhD student: There's a word limit and I don't know how to cut this paper any shorte...
Senior coauthor:

[amit @gravicle]

Design for humans vs. Synergy for middle managers

[Data Science - Professor X @datascience_jmu]

Please share: We are looking for excellent candidates to fill 4 (open) tenured university professorships at #CAIDAS, the upcoming Center for Artificial Intelligence in Data Science @Uni_WUE!

https://t.co/ecMHrzzjOS
#HCI #NLP #ML #MachineLearning #Mathematics #CS

[Konstantin Kobs @konstantinkobs]

@annargrs @emnlp2020 We haven't gotten any message even though our paper was accepted for Findings. Were they really sent out? The official @emnlp2020 account and their website does not mention this, either.

[Data Science - Professor X @datascience_jmu]

Our Discovery Challenge "ChAT - Chat Analytics for Twitch" will be presented next week at @ECMLPKDD 2020. Join us on Friday, Sep 18th, to discuss results and our exciting new Twitch data!

https://t.co/jpbh5tZ28b