{
  "id": 20383385,
  "title": "Learning the language of machine learning",
  "url": "https://urgent.news/2026/11/17/learning-the-language-of-machine-learning",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-17T05:00:00.000Z",
  "source": {
    "name": "MIT News Research",
    "slug": "mit-news-research",
    "url": "https://news.mit.edu/2026/learning-language-of-machine-learning-jacob-andreas-1117"
  },
  "original_language": "en",
  "account": "Jacob Andreas, a newly tenured associate professor, is working on developing techniques that enable artificial intelligence models to explain their learning process in a language humans can understand. This is a significant challenge in high-stakes scenarios where researchers desire to understand how these complex models work. Andreas, an associate professor in Electrical Engineering and Computer Science at MIT, aims to build intelligent systems that can learn from human guidance and communicate effectively with human collaborators.\n\nAndreas' fascination with language began in his childhood, reading voraciously and later immersing himself in computer programming. He pursued a career as an academic researcher, completing an undergraduate thesis project on machine translation, focusing on adapting existing statistical techniques into more powerful AI systems. This led to his studies in computational translation problems while he was a graduate student at the University of California, Berkeley.\n\nDuring this period, Andreas witnessed a seismic shift in the field of natural language processing, transitioning from statistical methods to deep learning techniques. This shift presented new challenges in making AI models more transparent and interpretable. Andreas and his team at MIT tackle these challenges by developing methods to make the training process more natural and efficient, enabling AI models to learn in a way more akin to human learning.\n\nIn addition to his work on interpretable machine learning, Andreas also explores interpretability in computer vision systems, further expanding the scope of his research. In 2019, he joined the MIT faculty, eager to collaborate with leading AI researchers while pushing the boundaries of modern machine learning. His group's work focuses on two main areas: developing natural language descriptions of AI models' internal states and making the training process more efficient and understandable for humans.",
  "summary": "By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.",
  "key_points": [
    "Jacob Andreas aims to make AI models explainable in human language.",
    "He transitioned from machine translation to deep learning in his research.",
    "Andreas focuses on natural language descriptions of AI models' internal states."
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}