Daniel Klotz (@ido87) header
Daniel Klotz (@ido87) profile picture

🏷️ ml + earth @ UFZ & Google 🔬 measuring what the models can and can’t.

@daniel_klotz@sigmoid.social
Joined August 2008
Photo by Daniel Klotz (@ido87)Photo by Daniel Klotz (@ido87)Photo by Daniel Klotz (@ido87)Photo by Daniel Klotz (@ido87)Photo by Daniel Klotz (@ido87)Photo by Daniel Klotz (@ido87)Photo by Daniel Klotz (@ido87)Photo by Daniel Klotz (@ido87)Photo by Daniel Klotz (@ido87)Photo by Daniel Klotz (@ido87)
Daniel Klotz retweeted
We’re excited to introduce TiRex — a pre-trained time series forecasting model based on an xLSTM architecture.
5
22
6
69
14,925
🌍 PhD in AI in Desert Geomorphology
🔎Join a @dfg_public funded project mapping desert pavements in Namibia using AI, remote sensing & fieldwork.
🗓️Start: 01 June Apply by: 07 Feb
📍Part-time (75%), fixed-term until 31 May 2028. @geobrenning 🔗 : jobs.uni-jena.de/jobposting/…
6
5
552
Daniel Klotz retweeted
The one and only @HochreiterSepp kicks of the 2nd day of @NeurIPSConf!
3
6
38
1,984
Deep Ensembles are widely used to improve the performance of Deep Learning models. But beware, they can have profound impact on group fairness ⚖️ We analyzed why it happens and what can be done about it 🧵👇
3
21
2
78
11,766
Daniel Klotz retweeted
Thrilled to share my second PhD paper in Environmental Research Letters! While most focus on spring preconditioning, our study reveals the impact of winter conditions on summer vegetation across the entire Northern Hemisphere. Read more: twtr.to/824hX
1
2
1
14
1,384
Inside WaterResearch - Issue 19: Dive into the performance of various models on the #GreatLakes—a region with complex transboundary datasets, diverse climatic and land use conditions, and immense socio-economic importance to both the USA and Canada. bit.ly/3y3izNM
1
4
15
882
Can we learn from the recent past to predict future climate change impacts using machine learning? We have created a new benchmark dataset designed to help answer this question, and you can take part in the challenge: kaggle.com/competitions/the-…
1
20
3
54
6,756
Daniel Klotz retweeted
Really excited to announce Energy-based Hopfield Boosting. 🎉 Hopfield Boosting advances the state-of-the-art in OOD detection. 🚀 Our new energy function allows a model to focus on hard training examples close to the decision boundary between in-distribution and OOD samples. 🧵
1
40
5
151
25,318
I am so excited that xLSTM is out. LSTM is close to my heart - for more than 30 years now. With xLSTM we close the gap to existing state-of-the-art LLMs. With NXAI we have started to build our own European LLMs. I am very proud of my team. arxiv.org/abs/2405.04517
45
358
76
1,819
276,755
I think neural network potentials are the most important scientific tool of the next decade. The ability to simulate systems at the molecular scale starting from nothing but quantum mechanics will be transformative for a vast range of problems throughout biology and chemistry 1/n
The Potential of #Neural Network Potentials A perspective from Timothy Duignan @TimothyDuignan @Griffith_Uni 🔓 Open access in ACS Physical Chemistry Au 👉 go.acs.org/8zg
9
93
13
615
134,705
xLSTM: Extended Long Short-Term Memory The famous LSTM nets are improved by exponential gates and matrix-memory with covariance update. Strong results on large-scale language modeling. P: arxiv.org/abs/2405.04517
5
49
4
205
47,539
Daniel Klotz retweeted
We are looking for your participation in the survey “AI/ML in the #Hydrology Community” – a joint initiative by #UNESCO #IHP & its EURO-FRIEND Project-3. Link to the survey: forms.office.com/e/5bkkBypXg… Please share widely! #hydrologyML #hydrologyAI #AIMLhydrology
3
26
5
41
14,031
Daniel Klotz retweeted
Nature research paper: Global prediction of extreme floods in ungauged watersheds go.nature.com/3Vrpzxj
1
18
1
50
21,999
Daniel Klotz retweeted
The Italian FlOod and Catchment Atlas (FOCA), a product of the #PNRR #Return Project, is out! Free data are on 631 basins are on Zenodo (doi.org/10.5281/zenodo.80607…). The full paper is free on ESSD (essd.copernicus.org/articles…). @Gii_Idraulica @ISPRA_Press @giornaleprociv @EGU_HS
7
2
26
3,315
Daniel Klotz retweeted
Large-scale global flood forecasting has been out of reach for a long time. In our Nature paper published today we show how breakthroughs in AI can close the gap & provide reliable flood predictions even in regions that previously lacked data. Learn more: goo.gle/49bRqVG
20
123
13
482
69,156
Daniel Klotz retweeted
Looks interesting 🤔
Discussion open for comments: Technical Note: The Divide and Measure Nonconformity doi.org/10.5194/hess-2024-59
1
1
10
1,376
Interested in a semantic memory for reinforcement learning? I was recently invited to a podcast talking about our #NeurIPS2023 paper: Semantic HELM (arxiv.org/abs/2306.09312). In case you are interested, you can stream the episode here: open.spotify.com/episode/4n2…
Excited to share our latest work on a semantic and interpretable memory module for RL! Complementary to recent developments in the realm of explainable AI, we focus on interpretability w.r.t. the memory of an agent. 1/n
1
15
23
2,051
Daniel Klotz retweeted
Our latest paper challenges the status quo in Unsupervised Learning for Combinatorial Optimization (CO) problems.💡 🔄 Unveiling the limitations of popular Mean Field methods, we showcase the power of Autoregressive Approximations of solving CO Problems on graphs. 🧵
1
25
2
53
13,521