We’re now onto the second month of 2023 and January certainly hasn't been a dull month! We’ve previously seen the importance of artificial intelligence in fields like medicine, the environment, and history, and the trend continues.
From helping to pollinate flowers to finding a beloved author’s lost play and more, we’ll dive into what happened in the world of AI in the past month.
In this article, we’ll take a look at a few headlines in AI, computer vision, and machine learning, including:
- Computer vision helps to locate king flowers on apple trees
- Machine learning spots eight potential technosignatures
- Plainsight and MarineSitu use computer vision to protect sea creatures
- AI uncovers unknown play by Spanish author in library archive
- AI helps discovery of super tight-binding antibodies
Penn State researchers developed a computer vision system that can locate and identify apple king flowers in blossom clusters on orchard trees. This first-of-its-kind study is an important first step to creating a robotic pollination system.
These blossoms grow in groups of four to six blooms on the branches, with the center one being known as the king flower. This is the first flower in the cluster to open and the one that grows the largest fruit.
Insect pollination isn't matching current demand, from both wild pollinators and domesticated honeybees, which has led scientists to develop alternative pollination methods. The study was conducted by Long He, assistant professor of agricultural and biological engineering at the College of Agricultural Sciences.
Xinyang Mu, a doctoral student in the Department of Agricultural Biological Engineering, used Mask R-CNN, a deep learning program that does pixel-level segmentation to detect objects partially obscured by other objects.
In their question to answer the question, “Are we alone in the universe?”, scientists used a new machine learning technique to find eight previously undetected ‘signals of interest’ from five nearby stars.
They applied an algorithm to previously studied data collected by the Green Bank Telescope in West Virginia. This was part of a campaign by the privately funded initiative Breakthrough Listen, which searches one million nearby stars, the Milky Way, and a hundred nearby galaxies for technologically advanced life.
Analyzing 150 terabytes of data from 820 nearby stars, the algorithm’s strength lies in organizing data from telescopes into categories to separate real signals from interference. The most successful algorithm used a combination of supervised learning and unsupervised learning, or semi-unsupervised learning.
This resulted in eight signals being found from five different stars between 30 to 90 lightyears from Earth.
Artificial intelligence company Plainsight partnered with MarineSitu, a software and hardware provider and spinoff from the University of Washington to help enable marine energy devices to coexist in harmony with aquatic life.
This involves an underwater camera for continuous monitoring and real-time processing of data, starting with tidal turbines. It's being taught about fish and their different types, and later on, it’ll learn about sizes and their optimal health.
This collaboration will help in understanding the ecosystem and potential negative impact on locations when looking for a tidal turbine site.
Artificial intelligence tech utilized to transcribe anonymous works at the archives of the National Library of Spain found a previously unknown play by author Felix Lope de Vega. The Baroque playwright wrote “La Francesa Laura”, or The Frenchwoman Laura, a few years before his death in 1635.
Researchers from both Vienna and Valladolid universities used the tech to transcribe 1,300 anonymous books and manuscripts, simultaneously saving years of work and trying to find their authors by checking the work against other writers' words.
The National Library stated that the words in “La Francesa Laura” were closely aligned with Lope’s, and not aligned with the other 350 playwrights that were part of this experiment.
Scientists from the University of California San Diego School of Medicine developed an artificial intelligence-based strategy to find high-affinity antibody drugs. In their study, researchers identified a new antibody that binds a major cancer target 17-fold tighter than what an existing antibody drug is able to do.
This discovery could help propel the development of new drugs to fight cancer and diseases like rheumatoid arthritis, as successful drugs need antibodies to tightly bind to their target. But in order to find these antibodies, scientists usually start with a known antibody amino acid sequence, using yeast or bacterial cells to produce a series of new antibodies with variations of the sequence.
Machine learning has helped accelerate and improve these efforts, as researchers can generate an initial library of around half a million possible antibody sequences and screen them for affinity to specific protein targets.
The team’s artificial intelligence model can report the certainty of each prediction, unlike other methods, helping researchers to rank antibodies and decide which ones are to be prioritized for drug development.