• AI KATANA
  • Posts
  • A Quant Firm Gets a Performance Boost From AI

A Quant Firm Gets a Performance Boost From AI

Also: Eight AI Startups Winning the Race for Tech Talent

Author

AI KATANA
July 04, 2023

Hi!

In the AI space, Robeco, a quantitative investment firm, reported a promising performance boost from machine learning tools in their investment process. The Biden-Harris Administration held a listening session with union leaders discussing the impacts and concerns of AI on labor and job markets. Researchers developed a deep learning model, TIGER, for precise control over gene expression with CRISPR technology. AI startups, known as the "Generative Eight," are attracting top talent from tech giants, with Midjourney releasing a new feature called Pan. A new programming language, Mojo, aims to speed up Python code performance specifically for AI developers. Lastly, Flywheel, a medical imaging data and AI platform provider, raised $54 million in Series D funding to consolidate growth and accelerate expansion into new markets.

Sliced perfectly:

  • 🧮 A Quant Firm Gets a Performance Boost From AI — At Least in Tests

  • 🏛️ White House Listening Session with Union Leaders on Advancing Responsible AI Innovation

  • 🧬 AI and CRISPR precisely control gene expression

  • 👨🏻‍💻 Eight AI Startups Winning the Race for Tech Talent

Quantitative investment firm Robeco, with $170 billion in assets under management, has been experimenting with machine learning (ML) tools to enhance its performance and has reported promising results. They integrated ML into their investment process for quantitative credit, starting with $5.5 billion in credit strategies. The results indicated a significant improvement in risk-adjusted returns for strategies using the value factor, with the information ratio (a measure of excess return per unit of risk) increasing from 1.42 to 1.83 for an investment-grade bond portfolio and from 0.99 to 1.75 for a high-yield bond portfolio. The firm's exploration of ML mirrors a broader trend among asset managers to harness AI for investment and distribution. The research stage for Robeco's ML approach, which took six months, has been completed, and they are now preparing for a software update to fully implement it. Looking ahead, the team plans to apply the same ML techniques to other factors such as momentum and size.

On June 30, senior Biden-Harris Administration officials held a listening session with union leaders to discuss the impacts of artificial intelligence (AI) on workers, unions, and the quality of jobs. Union leaders voiced concerns over the risks AI posed to job security, physical and mental health, privacy, and civil rights, but also recognized potential opportunities for AI to improve workers' lives with collaborative solutions involving unions and employers. The session, attended by officials from the White House National Economic Council, Office of Science and Technology Policy, and Office of the Vice President, emphasized the need for government and employers to work with unions to understand and mitigate these risks. Participants shared experiences of AI deployment in workplaces, noting both negative impacts, such as inaccuracies, increased workplace stress, and privacy invasion, and positive ones, such as reduced physical strain from repetitive or dangerous tasks. Union leaders agreed that AI was changing the skill sets required from workers and posed threats to creators' ownership of their intellectual property, with observations of AI being used to cut jobs.

Researchers from New York University, Columbia Engineering, and the New York Genome Center have developed a deep learning model called TIGER, enabling precise control over gene expression in conjunction with CRISPR technology. The model is designed to predict on-target and off-target activity of RNA-targeting CRISPR, a variant of the gene-editing tool which manipulates RNA rather than DNA. This tool could be used to treat diseases by reducing or eliminating the expression of specific genes, and is also valuable in the study and potential treatment of viral infections, as RNA is the primary genetic material in many viruses. The TIGER model is expected to play a crucial role in preventing unintended off-target effects of gene editing, thus furthering the development of CRISPR-based therapies.

The article from Lightspeed Venture Partners discusses the eight leading AI startups that are attracting top technical talent, referred to as the "Generative Eight". These companies are Tome, Character.ai, Anthropic, OpenAI, Hugging Face, Jasper, Stability AI, and Midjourney. Generative AI is rapidly altering the relationship between humans and technology, requiring unique technical skills and capabilities. The engineers adept at pushing the boundaries of today's AI technologies are often those with years of experience at large tech companies. However, these large companies are losing their talent to nimble, innovative AI startups that have AI at their core. The startups are competing for talent, with the Generative Eight leading the race. They've hired the highest caliber engineers and AI experts, found funding success, and achieved some level of product-market fit. This talent comes from leading tech companies across various sectors, including some employees who have moved on from OpenAI and Google. The tech talent at these companies also has a strong educational background, often from internationally recognized institutions.

🛠️ AI tools updates

Mojo is a new programming language designed specifically for AI developers, aiming to combine the usability of Python with the performance of C. This language, developed by Chris Lattner, is a strict superset of Python, ensuring full compatibility with the Python ecosystem while also developing as a first-class language. Mojo promises to speed up Python code by a factor of up to 3,500, depending on the hardware used, and offers systems-level optimizations that Python cannot. The development team plans to open-source Mojo, but a release date is yet to be announced. Despite the ongoing need for a solution to Python's performance issues in big data, machine learning, and AI, Mojo positions itself as a promising intermediary solution.

💵 Venture Capital updates

Flywheel, a Minneapolis-based company that provides a medical imaging data and AI platform, has successfully raised $54 million in Series D funding. The investment round was co-led by Novalis LifeSciences LLC and NVentures, NVIDIA’s venture capital arm, with Microsoft and several other entities also participating. The new funds will be used to consolidate Flywheel's growth in its main markets, namely public sector healthcare and pharmaceutical companies, and to speed up its expansion into other markets such as providers, payers, system integrators, and software companies. The company, headed by CEO Jim Olson, aims to power healthcare innovation by offering comprehensive solutions for efficient medical imaging data management, curation, and analysis, thus aiding rapid research and AI development.

🫡 Meme of the day

⭐️ Generative AI image of the day

Before you go, check out America's first law regulating AI bias 👩🏻‍⚖️