Hiroto, Kazuki, Nori, and Jungo will be lecturing about the basics of Mamba, a deep learning model that has recently gained traction as an alternative to transformers.

Noriyuki will give a talk about Kotoba's development of LLMs and our interest in the medical field at the meet-up co-hosted by MACNICA, Inc., Mitsubishi Estate, and Nvidia Japan. Kotoba will also host a booth to interact with Japanese medical companies.

One paper is presented at NeurIPS 2023, a top-tier international conference in AI and machine learning:

• RealTime QA: What's the Answer Right Now? (NeurIPS Datasets and Benchmarks Track).

Congratulations, Jungo and Keisuke!

Three papers are presented at EMNLP 2023, a top-tier international conference in AI and natural language processing:

• Do All Languages Cost the Same? Tokenization in the Era of Commercial Language Models (EMNLP Main Conference)

• Contrastive Learning-based Sentence Encoders Implicitly Weight Informative Words (EMNLP Findings)

• Batch Prompting: Efficient Inference with Large Language Model APIs (EMNLP Industry Track)

Congratulations Hiroto and Jungo!

Noriyuki provided a special virtual talk at the ITmedia Virtual Expo 2023 Fall. 8,000 people visited virtually. Thank you!

-> Event Website (JP)

Noriyuki was interviewed on TV Tokyo's 'Dawn of Gaia' (ガイアの夜明け) to discuss Kotoba Technologies' significant role in the Fugaku project and Kotoba's future vision as a startup.

-> Watch Video (JP)

Noriyuki provided comments on the CNBC article regarding the current status of LLM (Large Language Model) development in Japan

-> Full Article (EN)

Jungo's research on benchmarking GPT-4 for the Japanese medical bar exam was featured in an interview with Yomiuri, one of Japan's largest newspapers.

-> Full Article (JP)