Bloomberg trained an LLM from scratch on AWS (64 × 8 A100 40GB for 53 days). They constructed a 363 billion token dataset based on Bloomberg’s extensive data sources, perhaps the largest domain-specific dataset yet, augmented with 345 billion tokens from general-purpose datasets.

BloombergGPT outperforms similarly-sized open models on financial NLP tasks by significant margins — without sacrificing performance on general LLM benchmarks

Few-Shot Learning in NLP

Two recent papers on few-shot learning in NLP caught my eye: 1st on retrieval by Google Research and 2nd on classification by Intel and HuggingFace

Dai, Zhuyun, et al. “Promptagator: Few-shot Dense Retrieval From 8 Examples.” arXiv preprint arXiv:2209.11755 (2022).

we suggest to work on Few-shot Dense Retrieval, a setting where each task comes with a short description and a few examples. To amplify the power of a few examples, we propose Prompt-base Query Generation for Retriever (Promptagator), which leverages large language models (LLM) as a few-shot query generator, and creates task-specific retrievers based on the generated data.

Tunstall, Lewis, et al. “Efficient Few-Shot Learning Without Prompts.” arXiv preprint arXiv:2209.11055 (2022).

we propose SetFit (Sentence Transformer Fine-tuning), an efficient and prompt-free framework for few-shot fine-tuning of Sentence Transformers (ST). SetFit works by first fine-tuning a pretrained ST on a small number of text pairs, in a contrastive Siamese manner. 

Airbnb Search Papers

Grbovic, Mihajlo, and Haibin Cheng. “Real-time personalization using embeddings for search ranking at airbnb.” Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 2018. [blog 1, blog 2]

Haldar, Malay, et al. “Applying deep learning to airbnb search.” Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 2019.

Haldar, Malay, et al. “Improving deep learning for airbnb search.” Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 2020.

Abdool, Mustafa, et al. “Managing diversity in airbnb search.” Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 2020.

Embedding-based Search Retrieval Papers and Blogs

Papers to Start with

Some papers I really love:

  • Zhang Zhi, et al. “Bag of Freebies for Training Object Detection Neural Networks.” arXiv preprint arXiv:1902.04103 (2019).
  • Xie, Junyuan, et al. “Bag of Tricks for Image Classification with Convolutional Neural Networks.” arXiv preprint arXiv:1812.01187 (2018).
  • Howard, Jeremy, and Sebastian Ruder. “Universal language model fine-tuning for text classification.” Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Vol. 1. 2018.
  • Smith, Leslie N. “A disciplined approach to neural network hyper-parameters: Part 1–learning rate, batch size, momentum, and weight decay.” arXiv preprint arXiv:1803.09820 (2018).
  • Chahal, Karanbir, Manraj Singh Grover, and Kuntal Dey. “A Hitchhiker’s Guide On Distributed Training of Deep Neural Networks.” arXiv preprint arXiv:1810.11787 (2018).
  • Neishi, Masato, et al. “A bag of useful tricks for practical neural machine translation: Embedding layer initialization and large batch size.” Proceedings of the 4th Workshop on Asian Translation (WAT2017). 2017.
  • Joulin, Armand, et al. “Bag of tricks for efficient text classification.” arXiv preprint arXiv:1607.01759 (2016).
  • Covington, Paul, Jay Adams, and Emre Sargin. “Deep neural networks for youtube recommendations.” Proceedings of the 10th ACM Conference on Recommender Systems. ACM, 2016.
  • He, Xinran, et al. “Practical lessons from predicting clicks on ads at Facebook.” Proceedings of the Eighth International Workshop on Data Mining for Online Advertising. ACM, 2014.
  • McMahan, H. Brendan, et al. “Ad click prediction: a view from the trenches.” Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining. ACM, 2013.