Hacker News
- Neural network training makes beautiful fractals https://sohl-dickstein.github.io/2024/02/12/fractal.html 64 comments
- Deep physical neural networks trained with backpropagation https://www.nature.com/articles/s41586-021-04223-6 36 comments
- Show HN: I trained a recurrent neural network in JavaScript to draw dick doodles https://github.com/dickrnn/dickrnn.github.io/blob/master/README.md 2 comments
- Show HN: I trained a recurrent neural network to draw dick doodles https://dickrnn.github.io 3 comments
- One-Shot Training of Neural Networks Using Hypercube-Based Topological Coverings http://arxiv.org/abs/1901.02878v1 17 comments
- Researchers Introduce MERLIN For Training Despeckling Deep Neural Network-Based Models https://arxiv.org/pdf/2110.13148.pdf 2 comments artificial
- GPUs May Be Better, Not Just Faster, at Training Deep Neural Networks https://www.unite.ai/gpus-may-be-better-not-just-faster-at-training-deep-neural-networks/ 3 comments artificial
- Autodiff: Automatic Differentiation for Training Neural Networks on Embedded systems. https://github.com/mlaisd/autodiff-embs 3 comments embedded
- The Boundary of Neural Network Training is Fractal - across up to 10 orders of magnitude https://www.youtube.com/watch?v=4iDB9eTeBwY 3 comments artificial
- Watching a Model Train (Lisp, neural networks, Emacs. And low-tech) http://lisp-univ-etc.blogspot.com/2021/10/watching-model-train.html 5 comments lisp
- Resources on neural network training via meta-heuristic like GA/PSO? https://www.reddit.com/r/compsci/comments/fsfk3o/resources_on_neural_network_training_via/ 8 comments compsci
- Training a Neural Network? Start here! https://www.kaggle.com/lavanyashukla01/training-a-neural-network-start-here 3 comments datascience
- Training a Neural Network using WebAssembly and Rust https://juggernaut.rs/demo/ 13 comments rust
- Training a neural network to rhyme. https://www.reddit.com/r/artificial/comments/6rqac1/training_a_neural_network_to_rhyme/ 4 comments artificial
- How to train an artificial neural network to play Diablo 2 using visual input http://stackoverflow.com/questions/6542274/how-to-train-an-artificial-neural-network-to-play-diablo-2-using-visual-input 7 comments programming
- Using a Neural Network to Train a Ruby Twitter bot http://www.fullstackfest.com/agenda/skynet-for-beginners-using-a-neural-network-to-train-a-ruby-twitter-bot 4 comments ruby
- Training a Neural Network to Paint an Image http://synaptic.juancazala.com/paint.html 5 comments javascript
- Training vanilla deep neural networks to classify numbers as being divisible by another number https://github.com/bigstronkcodeman/Deep-Neural-Network/blob/master/Neural.py 8 comments learnmachinelearning
- Another Neural Network Has Been Trained to Recognize Skin Cancer https://evolvera.tech/2021/10/30/another-neural-network-has-been-trained-to-recognize-skin-cancer/ 20 comments science
- Archaeologists train a neural network to sort pottery fragments for them https://arstechnica.com/science/2021/05/archaeologists-train-a-neural-network-to-sort-pottery-fragments-for-them/ 4 comments history
- [R] Carbon Emissions and Large Neural Network Training (Google Brain and Berkeley paper with Jeff Dean as the last author) https://arxiv.org/abs/2104.10350 5 comments machinelearning
- Intel and Baidu partner on Nervana Neural Network AI training processor https://techcrunch.com/2019/07/02/intel-and-baidu-partner-on-nirvana-neural-network-ai-training-processor/ 6 comments intel
- Neural Network Training Set Representation https://www.reddit.com/r/learnprogramming/comments/9loniy/neural_network_training_set_representation/ 3 comments learnprogramming
- This website allows you to ask the details of a song that you would like and an AI that was trained through neural networks creates a unique song right at the click of a button for free ! https://www.jukedeck.com/?invite=4282b4688f156da9c46059b15fb9a776917a9f58d127a46b1db4cfc35be2692d 13 comments artificial
- Stochastic gradient descent: from noisy gradients in millions of dimensions for neural network training - how to go to 2nd order methods? http://ruder.io/optimizing-gradient-descent/index.html 26 comments math
- New patent for "training action selection neural networks using look-ahead". Should we be worried? https://patentscope2.wipo.int/search/en/detail.jsf;jsessionid=A1D2B1253CB3B5C7A67A1BB18C25E86C?amp%3BmaxRec=3424245&%3Boffice=&%3BprevFilter=%26fq%3DOF%3AWO&%3BqueryString=&%3BrecNum=1809&%3BsortOption=Pub+Date+Desc&docId=WO2018215665 5 comments deeplearning
- Speak to the hive mind - query a deep neural network trained on reddit comments! http://comment.ai/ 37 comments programming
- I trained a neural network to generate GitHub repo names - the results are pretty amusing http://colinmorris.github.io/rbm/repos/ 141 comments programming
- [R] NoProp: Training neural networks without back-propagation or forward-propagation https://arxiv.org/pdf/2503.24322 31 comments machinelearning
- DALL·E: Creating Images from Text: OpenAI trained a neural network called DALL·E that creates images from text captions for a wide range of concepts expressible in natural language. https://openai.com/blog/dall-e/ 17 comments artificial
- [P] A self-driving car using Nvidia Jetson Nano, with movement controlled by a pre-trained convolution neural network (CNN) written in Taichi https://github.com/houkensjtu/taichi-hackathon-akinasan 21 comments machinelearning
- [R] How To Train a Neural Network with Less GPU Memory: Reversible Residual Networks Review https://www.opencv.ai/blog/train-neural-network-reversible-residual-networks 5 comments machinelearning
- Effects of Music Training on Inhibitory Control and Associated Neural Networks in School-Aged Children: A Longitudinal Study https://doi.org/10.3389/fnins.2019.01080 4 comments science
- Elon Musk Says Tesla Developing Neural Network Training Computer for Full Self-Driving https://www.ibtimes.sg/elon-musk-says-tesla-developing-neural-network-training-computer-full-self-driving-50129 323 comments technews
- Indoor solar cells with 34% efficiency can power IoT devices and machine learning, converting indoor light into structured information. The collection of 2.72x10^15 photons (0.947mJ) is needed to compute one inference of a pre-trained artificial neural network for MNIST image classification. https://pubs.rsc.org/en/content/articlelanding/2020/sc/c9sc06145b 6 comments science
- Researchers have trained a neural network to recognize rock samples in core box images efficiently. It speeded up the analysis process up to 20 times and made it possible to automate the description of rock samples. https://www.sciencedirect.com/science/article/pii/S0098300422000619 2 comments science
- Researchers Build AI That Builds AI - Given a new, untrained deep neural network designed for some task, the hypernetwork predicts the parameters for the new network in fractions of a second, and in theory could make training unnecessary. https://www.quantamagazine.org/researchers-build-ai-that-builds-ai-20220125/ 5 comments artificial
- [R] Check the blog that introduces our recent #NeurIPS2021 work on nonconvex stochastic optimization! We propose stochastic Anderson mixing with theoretical guarantees and promising results in training neural networks. https://thumtblog.github.io/2021/10/30/stochastic-anderson-mixing/ 2 comments machinelearning
- [R] Impact of GPU uncertainty on the training of predictive deep neural networks: When training a predictive neural net using only CPUs, the learning error is higher than when using GPUs, suggesting that GPUs plays a different role in the learning process than just increasing computational speed. https://arxiv.org/abs/2109.01451 33 comments machinelearning
- A new theory inspired by neural networks suggests that the weirdness of dreams is essential to their purpose. Neural networks can become "overfitted" to training data, reducing their use. The theory proposes that weird dreams act to stop our brains becoming "overfitted" to boring everyday life. https://www.technologynetworks.com/neuroscience/news/the-weirdness-of-our-dreams-could-explain-their-function-suggests-new-theory-348797 280 comments science