-
Gwern.net newsletter (Substack subscription page)
-
August 2021 News
-
‘newsletter’ directory
-
Changelog
-
Gwern Branwen Creating Essays on Gwern.net
-
The Math of Hunting Lions
-
[SIGGRAPH Asia 2021] DeepVecFont: Synthesizing High-Quality Vector Fonts via Dual-Modality Learning
-
How Does AI Improve Human Decision-Making? Evidence from the AI-Powered Go Program
-
Welcome to State of AI Report 2021
-
Achieving Human Parity on Visual Question Answering
-
Transformers are Meta-Reinforcement Learners
-
PFNs: Transformers Can Do Bayesian Inference
-
Generate Images from Texts. In Russian
-
The Greatest Remaining Hits
-
https://ooo.ghostbows.ooo/about/
-
Making This Album With AI ‘Felt like Wandering in an Enormous Labyrinth’: Shadow Planet Is the Result of a Three-Way Collaboration between Humans and AI
-
https://ooo.ghostbows.ooo/story/
-
https://github.com/EdZ543/This-Catgirl-Does-Not-Exist
-
ML Scaling subreddit
-
Collaborating with Humans without Human Data
-
A Recipe For Arbitrary Text Style Transfer with Large Language Models
-
Story Centaur: Large Language Model Few Shot Learning As a Creative Writing Tool
-
GPT-3 Creative Fiction § Literary Parodies
[Transclude the forward-link's context]
-
A Review of Text Style Transfer using Deep Learning
-
The Bitter Lesson
-
Symbolic Knowledge Distillation: from General Language Models to Commonsense Models
-
MAE: Masked Autoencoders Are Scalable Vision Learners
-
Visible Thoughts Project and Bounty Announcement
-
LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairs
-
CoAtNet: Marrying Convolution and Attention for All Data Sizes
-
Turing Bletchley: A Universal Image Language Representation Model by Microsoft
-
Vector-quantized Image Modeling with Improved VQGAN
-
Autoregressive Latent Video Prediction with High-Fidelity Image Generator
-
Effect of scale on catastrophic forgetting in neural networks
-
On the Opportunities and Risks of Foundation Models
-
A Universal Law of Robustness via Isoperimetry
-
https://x.com/SebastienBubeck/status/1402645428504461319
-
A Universal Law of Robustness
-
scaling-hypothesis#blessings-of-scale
[Transclude the forward-link's context]
-
Why Robust Generalization in Deep Learning is Difficult: Perspective of Expressive Power
-
On the Predictability of Pruning Across Scales
-
‘NN sparsity’ directory
-
Sparse is Enough in Scaling Transformers
-
What Are Bayesian Neural Network Posteriors Really Like?
-
Greedy Hierarchical Variational Autoencoders (GHVAEs) for Large-Scale Video Prediction
-
https://wandb.ai/dalle-mini/dalle-mini/reports/DALL-E-mini--Vmlldzo4NjIxODA
-
DALL·E 1: Creating Images from Text: We’ve trained a neural network called DALL·E that creates images from text captions for a wide range of concepts expressible in natural language
-
Multimodal Few-Shot Learning with Frozen Language Models
-
AudioCLIP: Extending CLIP to Image, Text and Audio
-
Google Demonstrates Leading Performance in Latest MLPerf Benchmarks
-
Time-Aware Language Models as Temporal Knowledge Bases
-
Genetic risk factors have a substantial impact on healthy life years
-
Novel disease associations with schizophrenia genetic risk revealed in ~400,000 UK Biobank participants
-
Does parental education influence child educational outcomes? A developmental analysis in a full-population sample and adoptee design
-
Using genes to explore the effects of cognitive and non-cognitive skills on education and labor market outcomes
-
The contribution of additive genetic variation to personality variation: heritability of personality
-
Uncovering the Genetic Architecture of Broad Antisocial Behavior through a Genome-Wide Association Study Meta-analysis.
-
Genome-wide association analyses of individual differences in quantitatively assessed reading-related and language-related skills in up to 34,000 people
-
Identifying a Living Great-Grandson of the Lakota Sioux Leader Tatanka Iyotake (Sitting Bull)
-
Global Biobank Meta-analysis Initiative: powering genetic discovery across human diseases
-
https://academic.oup.com/nsr/article/8/6/nwaa175/5893295
-
Facultative Parthenogenesis in California Condors
-
Sigmoids behaving badly: why they usually cannot predict the future as well as they seem to promise
-
Maths Writer/cowritter Needed: How You Can’t Distinguish Early Exponential from Early Sigmoid
-
Why Sigmoids Are so Hard to Predict
-
Forecasting S-Curves Is Hard
-
Predicting the Future With Data+logistic Regression
-
Inconsistency in Conference Peer Review: Revisiting the 2014 NeurIPS Experiment
-
Gender Identity, Coworking Spouses and Relative Income within Households
-
The Razor Blade in the Apple: The Social Construction of Urban Legends
-
Obesity of politicians and corruption in post-Soviet countries
-
Rule Enforcement Without Visible Means: Christmas Gift Giving in Middletown
-
https://harpers.org/archive/2021/04/lost-in-thought-psychological-risks-of-meditation/
-
Remembering immunity: Neuronal ensembles in the insular cortex encode and retrieve specific immune responses
-
The Brain Can Recall and Reawaken Past Immune Responses: The Brain Not Only Helps to Regulate Immune Responses, but Also Stores and Retrieves ‘Memories’ of Them
-
Erosion of the Epigenetic Landscape and Loss of Cellular Identity as a Cause of Aging in Mammals
-
Making sense of the ageing methylome
-
Blood-based epigenome-wide analyses of cognitive abilities
-
A collective analysis of lifespan-extending compounds in diverse model organisms, and of species whose lifespan can be extended the most by the application of compounds
-
A New Way to Be Mad
-
Echolocating bats rely on an innate speed-of-sound reference
-
Attractor and integrator networks in the brain
-
Reality shifting: psychological features of an emergent online daydreaming culture
-
Nothing Ventured, Nothing Gained: Parasite Infection is Associated with Entrepreneurial Initiation, Engagement, and Performance
-
Escape of hair follicle stem cells causes stem cell exhaustion during aging
-
Synthetic fat from petroleum as a resilient food for global catastrophes: Preliminary techno-economic assessment and technology roadmap
-
David Denkenberger on Using Paper Mills and Seaweed to Feed Everyone in a Catastrophe, Ft Sahil Shah
-
2005: Shades of Doom
-
Toshi Omagari
-
ToughSF: Anti-Orbit Laser Submarines
-
Is this the simplest (and most surprising) sorting algorithm ever?
-
How inefficient can a sort algorithm be?
-
Coolex: The coolest way to generate combinations
-
Choices Are Really Bad
-
The science of cycology: Failures to understand how everyday objects work
-
Donors Vastly Underestimate Differences in Charities’ Effectiveness
-
https://www.uliwestphal.de/elephas-anthropogenus/index.html
-
Elephas anthropogenus
-
Distinguishing Real Vs. Fake Tiger Penises [Identification Guides for Wildlife Law Enforcement #6]
-
https://www.reddit.com/r/TOUHOUMUSIC/search/?q=author%3Agwern&sort=new&restrict_sr=on&t=all
-
https://www.reddit.com/r/Vocaloid/search/?q=author%3Agwern&sort=new&restrict_sr=on&t=all
-