Jörn Hees
582 posts
Jörn Hees
@joernhees
Machine Learning, Deep Learning, Data Mining, AI, Semantic Web, Linked Data guy... teaching machines to associate like humans
Germany Katılım Mayıs 2011
359 Takip Edilen370 Takipçiler
I just stumbled upon a call for journal contributions with the sentence "For each accepted paper in JEA, the author will receive a remuneration of 200 dollars". What? I mean, what? anser.press/index.php/JEA/…
English
Jörn Hees retweetledi
Excited to share our latest survey paper on image super-resolution🖼️🔍, accepted by #TPAMI: ieeexplore.ieee.org/document/10041…
@rave78 @spalaciob @joernhees @stfrolov @ComputerSociety
#ImageSuperResolution #IEEE #IEEECS #ComputerVision #MachineLearning
English
@vodafone_de Supi, wenn man in der Telefonschleife dann “anderes Thema” sagt, dann kommt “ich habe das Thema nicht verstanden, wenn Sie sich entschieden haben, rufen Sie uns wieder an… Klick” 🤣
Deutsch
@joernhees Am besten bitte den Vorfall telefonisch unter 08001721212 melden, damit meine Kolleg:innen prüfen können, ob der Anruf von uns kam 💪
Deutsch
@joernhees Das hört sich ganz stark nach einem Phishing-Anruf an. Woran Du erkennst, dass wir anrufen, erfährst Du in dem Bericht 💪 vodafone.de/featured/servi…
Deutsch
@ulmerleben @cestmauvaisca Haben mich heute in der Gegenrichtung auch gleich mehrfach geärgert… #diebahnsteht
Deutsch
@RubenVerborgh @danbri When downtime of your server becomes exam relevant 🤣
English
Jörn Hees retweetledi
The fifth hyperfactorial: 5⁵ × 4⁴ × 3³ × 2² × 1¹ = 86,400,000 milliseconds is exactly 1 day 🤓
English
So, it’s official… I’m coming back 😊 h-brs.de/de/news/joern-… @h_bonnrheinsieg #H_BRS @fs_inf #DataScience
English
As it turned out, solving the problem of learning hierarchical representations and complex functional dependencies was a much more important issue than being able to perform accurate probabilistic inference with shallow models.
9/N
English
Researchers in speech recognition, computer vision, and natural language processing in the 2000s were obsessed with accurate representations of uncertainty.
1/N
English
@janmartinkeil @kidehen @dbpedia @github Count these things - 200 OK, result is 42… - did you count them all? - I mean it’s 200 OK, what do you think? - Cool, i guess 42 might be a reliable count then, but can you count again? - 200 OK, result is 1337 - 🤦♂️ this is my life now?
English
Im Restaurant:
- "Wäre super, wenn Sie draußen ein paar Fahrradständer hätten."
- "Dafür haben wir leider keinen Platz."
- "Sie haben da draußen etwa 50 Parkplätze, da könnten Sie doch 2-3 dafür nutzen?"
- "Aber wo sollen denn dann die Autos hin?"
#verkehrswende #Mannheim
Deutsch
What's an explanation? What's an interpretation? Have a look at this nice video by Sebastian from our group #XAI #explainability #MachineLearning #DeepLearning
Sebastian P@spalaciob
Confused about what "explanation" or "interpretation" mean when reading about #XAI? We introduce the XAI Handbook: a framework for exactly this! We'll be presenting it on R-PR&MI @ICCV_2021. In the meantime, have a look at the video #ICCV2021 #ICCV21video youtu.be/5tIkL9p3heA
English
how do i quickly share files between to macbooks? (by quickly i mean fast and reliable transfer. need to transfers hundreds of gb. airdrop failed miserably at this. and somehow using the ssh server only gives me ~20mb/sec)
English
I’ve never had someone else explain one of our papers on YouTube... not gonna lie, i got beer and chips and watched it on the couch, and got even more hyped about our research ;). Good job @gordic_aleksa, made our days 👍
Aleksa Gordić (水平问题)@gordic_aleksa
[🎥🔥: new video - AudioCLIP paper explained!] Extending @OpenAI's CLIP (contrastive language-image pre-training) model to support the audio modality and achieving SOTA results on sound classification. YT: youtu.be/3SLQVh9ABDM @rave78
English
Jörn Hees retweetledi
[🎥🔥: new video - AudioCLIP paper explained!]
Extending @OpenAI's CLIP (contrastive language-image pre-training) model to support the audio modality and achieving SOTA results on sound classification.
YT: youtu.be/3SLQVh9ABDM
@rave78
YouTube
English
Jörn Hees retweetledi
AudioCLIP: Extending CLIP to Image, Text and Audio⋆
pdf: arxiv.org/pdf/2106.13043…
abs: arxiv.org/abs/2106.13043
achieves new sota results in the ESC task, out-performing other approaches by reaching accuracies of 90.07 % on the UrbanSound8K and 97.15 % on the ESC-50 datasets
English