Newsletter #56 - Autonomous tractors and trucks

Welcome to Nural's newsletter where you will find a compilation of articles, news and cool companies, all focusing on how AI is being used to tackle global grand challenges.

Our aim is to make sure that you are always up to date with the most important developments in this fast-moving field.

We now have a Jobs section, currently featuring an exciting data scientist role at startup AxionRay
Reach out to advertise your own tech roles!

Packed inside we have

An autonomous tractor controlled from a smartphone;
machine learning used to develop both cancer treatments and new materials;
and a video of an 80 mile, autonomous journey by a truck on public roads.

If you would like to support our continued work from £1 then click here!

Graham Lane & Marcel Hedman

Key Recent Developments

How John Deere created its autonomous tractor

What: U.S. tractor company John Dere demoed an autonomous tractor at the recent Consumer Electronics Show, announcing it will be available by the end of 2022. The tractor can autonomously navigate, plow, sow and avoid obstacles. It can also identify and kill unwanted weeds with a precision spray of herbicide which is claimed to reduce chemical usage by up to 80%. The tractor operates with a deep neural network and can be controlled from a smartphone.
Key Takeaways: The deployment of autonomous vehicles is progressing rapidly in farming and truck transport (reported elsewhere in this newsletter) since the operating environment is relatively simple. Some commentators are concerned that John Dere could be the “Facebook of farming” making farmers dependent on the company and exploiting data “harvested” by the tractors

Machine Learning helps repair genetic damage

What: Using ML, a team of researchers have discovered how to repair genetic damage to DNA. Chemotherapy works by causing DNA lesions, which cause cancer cells to die. Knowing how DNA lesions are repaired will shed light on cancer treatments. The researchers processed thousands of images of cells after genetic damage, introducing 300 different proteins to study what role they played and discovered nine new proteins that are involved in cell repair.
Key Takeaways: The large volume of data generated from images has been a limiting factor in studying DNA repair. The team mainly benefited from the use of ML to process high volumes of data more quickly. This included relatively simple, unsupervised K-means clustering that yielded new information.
Paper: Assessing kinetics and recruitment of DNA repair factors using high content screens

Machine learning used to predict synthesis of complex novel materials

What: A research team has applied machine learning to guide the synthesis of new nanomaterials, eliminating barriers associated with materials discovery. The team created a huge library of nano particles made of combinations of any of the 118 elements in the periodic table. This data was used to train an ML model. The model predicted entirely novel compositions of four, five and six elements that would result in a certain structural feature. It made 19 predictions, of which 18 were correct.
Key Takeaways: It is hoped this approach may drive discoveries across areas critical to the future, including plastic upcycling, solar cells, superconductors, and qubits. The team is using the approach to find green catalysts critical to fuelling processes in clean energy, automotive and chemical industries.
Paper: Machine learning–accelerated design and synthesis of polyelemental heterostructures

AI Ethics

🚀 Online public discourse on artificial intelligence and ethics in China: context, content, and implications

The article argues that the general public in China has a sense of anxiety towards AI rather than the widely held view that Chinese people are positive towards digital technologies.

🚀 China leads the world in computer vision surveillance research

The paper argues that China produces ‘a disproportionate share’ of research about core AI-related surveillance technologies.

🚀 Data Statements for Natural Language Processing: Toward mitigating system bias and enabling better science

A discussion of the idea that "data statements" (meta data and a description of datasets) will help practitioners to use data responsibly.

Cool companies found this week

Building automation

PassiveLogic - creates “digital twins” of buildings and land to enable customers to design, build, operate, and manage buildings and building controls. Has raised $34 million in round B funding.

Synthetic data

Mostly.ai - generates safe, realistic, smart synthetic data which it claims is the basis for fairer, responsible AI. The company has raised $25 million in round B funding.

Drug discovery

Protai - is developing a protein-centered, AI-powered drug discovery platform. It has raised $8 million in seed funding.

And finally ... whoever thought that your home assistant would be a robotic table.

AI/ML must knows

Foundation Models - any model trained on broad data at scale that can be fine-tuned to a wide range of downstream tasks. Examples include BERT and GPT-3. (See also Transfer Learning)
Few shot learning - Supervised learning using only a small dataset to master the task.
Transfer Learning - Reusing parts or all of a model designed for one task on a new task with the aim of reducing training time and improving performance.
Generative adversarial network - Generative models that create new data instances that resemble your training data. They can be used to generate fake images.
Deep Learning - Deep learning is a form of machine learning based on artificial neural networks.

Best,

Marcel Hedman
Nural Research Founder
www.nural.cc

If this has been interesting, share it with a friend who will find it equally valuable. If you are not already a subscriber, then subscribe here.

If you are enjoying this content and would like to support the work financially then you can amend your plan here from £1/month!