Transformers learn in-context by gradient descent Oswald et al.: https://
arxiv.org/abs/2212.07677 #MachineLearning #DeepLearning #ArtificialIntelligence
MACHINE LEARNING
-
Transformers Learn In-Context Through Gradient Descent
By
–
-
Major AI Conference Submission Deadlines January 2026
By
–
Some venues that come to mind include COLT, ICML, FAccT, FORC, and EC. These are mostly late January deadlines though.
-
Analysis of False Positives in GPT Plagiarism Detection
By
–
Follow-up to GPT "Plagiarism" Detector: False positives are common. With a piece of 100% original human writing, and the prediction came back as 84% Fake. It appears that syntax (sentence structure) influences the output quite a bit.
-
Early Style Transfer Research Shared from Fast.ai Course
By
–
Here's some of his earliest research that he shared with us shortly after the http://
fast.ai course wrapped up in 2017: https://
towardsdatascience.com/experiments-on
-different-loss-configurations-for-style-transfer-7e3147eda55e
… -
CALM: Dynamic Computational Effort for Language Models
By
–
Presenting Confident Adaptive Language Modeling (CALM), a novel method that allows language models to dynamically modify computational effort when generating text. Learn how CALM can accelerate text generation while preserving output quality → https://t.co/Nm7yyT8sMA pic.twitter.com/MpuaPBzDMU
— Google AI (@GoogleAI) 16 décembre 2022Presenting Confident Adaptive Language Modeling (CALM), a novel method that allows language models to dynamically modify computational effort when generating text. Learn how CALM can accelerate text generation while preserving output quality → https://
goo.gle/3HJKzbM -
ROSCOE: New Metrics Suite for Evaluating Step-by-Step Reasoning
By
–
ROSCOE is a first-of-its-kind suite of metrics for scoring step-by-step reasoning. By publishing this study we hope to provide a foundation that enables scalable systematic evaluation and benchmarking of new language models. See the paper on arXiv https://
arxiv.org/abs/2212.07919 -
Ludwig AI Maintainer Discusses Declarative ML Approaches
By
–
ICYMI: @ludwig_ai maintainer Justin Zhao joined @DataTalksClub to discuss all things declarative #ML. Topics covered:
• State of ML and #AutoML
• Overview of declarative approaches
• Live Ludwig demo
• Rapid config-driven iteration https://
pbase.ai/3HFNSAI #deeplearning -
LangChain Updates: Tokenizer, Document Reduction, Better Embeddings
By
–
TL;DR: The new tokenizer allows easy estimation of token usage The `collapse` step allows for map-reducing over large number of documents Better embeddings allows for better searching over those documents https://
github.com/hwchase17/lang
chain
… -
Collapse Step for Map-Reduce Chain Scaling
By
–
Adding a `collapse` step into the map-reduce chain This recursively collapses documents into smaller and smaller chunks, allowing you to scale to arbitrarily many documents This was a @johnjnay special – see his thread for more
-
OpenAI Releases New Embedding Model for Semantic Search
By
–
@OpenAI released a new embedding model (`text-embedding-ada-002`) This makes it possible to more easily do semantic search over collections of documents This is now the default OpenAI embedding model in LangChain