HN Explorer

groar

↗ HN profile [ 149 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. GPT Image 1 API is out (platform.openai.com)
    6 points 1
  2. Scientists detect signature of life on a distant planet, study suggests (cnn.com)
    1 points 0
  3. Codestral 25.01 by Mistral.ai (mistral.ai)
    1 points 0
  4. 1 points 2
  5. A Strong Correspondance Between Bayesian Networks and Linear Logic Proof-Nets (arxiv.org)
    1 points 0
  6. FlyAI – Injecting Responses of a Living Fly into Decision Making (arxiv.org)
    1 points 0
  7. Google proves Fault-Tolerant Quantum Computing is possible (arxiv.org)
    2 points 0
  8. SpikingBERT: Distilling Bert to Train Spiking Language Models (arxiv.org)
    2 points 0
  9. The Extractor: a drop-in replacement for self attention in Transformers (arxiv.org)
    1 points 0
  10. Super insects are devouring the forests (nautil.us)
    2 points 0
  11. New Google big language model Flan-T5 available on HuggingFace (huggingface.co)
    2 points 0
  12. SetFit: Efficient Few-Shot Learning Without Prompts (arxiv.org)
    1 points 0
  13. Neural Networks Are Decision Trees (arxiv.org)
    4 points 2
  14. YOLOv7: Trainable Bag-of-Freebies (arxiv.org)
    92 points 31
  15. Reconstructing Training Data from Trained Neural Networks (arxiv.org)
    1 points 0
  16. A Robust Backpropagation-Free Framework for Images (arxiv.org)
    2 points 0
  17. Deep physical neural networks trained with backpropagation (nature.com)
    108 points 37
  18. Can the brain use waves to solve planning problems? (arxiv.org)
    1 points 0
  19. Origins of the “OK” Word (en.wikipedia.org)
    17 points 1
  20. Neural Teleportation (arxiv.org)
    3 points 0
  21. Using Tree Neural Networks in Proof-Assistant HOL4 (arxiv.org)
    15 points 0
  22. Bitcoin crossed the $10k mark again (cnbc.com)
    4 points 0
  23. YOLOv4: Optimal Speed and Accuracy of Object Detection (arxiv.org)
    234 points 59
  24. The Cost of Training NLP Models: A Concise Overview (arxiv.org)
    2 points 0
  25. Cleaning algorithm finds 20% of errors in major image recognition datasets (deepomatic.com)
    221 points 81