PPaperPicks

Sepp Hochreiter

Johannes Kepler University of Linz, Austria

22 papers at tracked venues · 21 at CORE A* · active 20242025

Venues

Frequent coauthors

Papers

  1. A Large Recurrent Action Model: xLSTM enables Fast Inference for Robotics Tasks
  2. Bio-xLSTM: Generative modeling, representation and in-context learning of biological and chemical sequences
  3. Comparison Visual Instruction Tuning
  4. FlashRNN: I/O-Aware Optimization of Traditional RNNs on modern hardware
  5. Geometry-Informed Neural Networks
  6. Improving Uncertainty Estimation through Semantically Diverse Language Generation
  7. MIM-Refiner: A Contrastive Learning Boost from Intermediate Pre-Trained Masked Image Modeling Representations
  8. On Information-Theoretic Measures of Predictive Uncertainty
  9. Parameter Efficient Fine-tuning via Explained Variance Adaptation
  10. Rethinking Losses for Diffusion Bridge Samplers
  11. Scalable Discrete Diffusion Samplers: Combinatorial Optimization and Statistical Physics
  12. The Disparate Benefits of Deep Ensembles
  13. TiRex: Zero-Shot Forecasting Across Long and Short Horizons with Enhanced In-Context Learning
  14. Tiled Flash Linear Attention: More Efficient Linear RNN and xLSTM Kernels
  15. Vision-LSTM: xLSTM as Generic Vision Backbone
  16. pLSTM: parallelizable Linear Source Transition Mark networks
  17. xLSTM 7B: A Recurrent LLM for Fast and Efficient Inference
  18. A Diffusion Model Framework for Unsupervised Neural Combinatorial Optimization
  19. Contrastive Tuning: A Little Help to Make Masked Autoencoders Forget
  20. Energy-based Hopfield Boosting for Out-of-Distribution Detection
  21. Overcoming Saturation in Density Ratio Estimation by Iterated Regularization
  22. xLSTM: Extended Long Short-Term Memory