John X. Morris

Career

Morris took part in a Google Brain research program and later worked as a part-time student researcher at Meta's FAIR lab. Morris's site highlights research on text embeddings, including a training method that incorporates context at test time (2024) and methods for inverting text embeddings (2023).¹

Research

Morris's research focuses on text embeddings and language models. With Volodymyr Kuleshov, Vitaly Shmatikov and Alexander M. Rush, Morris showed in 2023 that text embeddings can be inverted to recover the original text, and with Rush introduced Contextual Document Embeddings (2024), which incorporate context from neighboring documents.⁵⁴¹

Morris is the second author, after Zach Nussbaum, of the Nomic Embed technical report on a reproducible long-context text embedding model. Earlier, Morris was first author of TextAttack (2020), a Python framework for adversarial attacks and data augmentation in NLP.²⁶

Work and contributions

• nomic-embed-text-v1.5 Open weights

  Second author of "Nomic Embed: Training a Reproducible Long Context Text Embedder", the technical report cited in the nomic-embed-text-v1.5 model card.²³

• TextAttack Open-source code

  First author of the TextAttack paper (2020), a framework for adversarial attacks, data augmentation and adversarial training in NLP, which Morris's site says Morris created in 2019 and 2020.⁶¹

Timeline

1. Apr 29, 2020Posted the TextAttack framework paper as first author.⁶
2. Oct 10, 2023Posted "Text Embeddings Reveal (Almost) As Much As Text" as first author.⁵
3. Feb 2, 2024Co-authored the Nomic Embed technical report as second author.²
4. Oct 3, 2024Posted Contextual Document Embeddings with Alexander M. Rush.⁴

Documented roles

The sources do not confirm a current role; these are the roles they document.

• Part-time student researcher, Meta (FAIR)¹
• Research program participant at Google Brain, Google Brain¹

Companies in their history

• Meta 106 people in the wiki
• Google 121 people in the wiki

Related people

• Adina WilliamsConnected to Meta
• Albert JiangConnected to Meta
• Jesse DodgeConnected to Meta
• Armand JoulinConnected to Meta
• Timothée LacroixConnected to Meta
• Leo GaoLanguage models · Datasets and data
• Sheshansh AgrawalLanguage models · Datasets and data
• Rohan TaoriLanguage models · Datasets and data
• Hynek KydlíčekDatasets and data · Language models
• Oren EtzioniDatasets and data · Language models