This week’s list of top data news highlights covers March 23, 2024 to March 29, 2024 and includes articles on using AI to identify birds and using drones to fight forest fires.
Orchard Robotics, a U.S.-based agricultural startup, has developed a sensing tool that attaches to existing farm equipment, such as tractors, to collect data on apple crops. The system’s cameras capture up to 100 images per second, recording information about every crop they pass, while the Orchard’s software utilizes AI to build detailed maps of the crop’s success rate. This approach provides farmers with actionable insights to optimize their yield and crop management practices.
2. Predicting Immune System Responses
Arizona State University researchers have created an AI-powered learning tool that can forecast how a person’s immune system will react to foreign cells. This tool identifies the distinct set of proteins that are unique to each individual and uses machine learning to predict whether the immune system will detect components of harmful viruses and cancers.
3. Improving Economic Transparency
The city of Tampa, Florida has launched a data dashboard that provides residents with information on 47 economic indicators. The data comes from nine different sources, seven of which are private datasets and two of which are publicly available datasets. The data shows Tampa consumer spending in the last three months of 2023 was up 37 percent from the same quarter in 2022.
Researchers at Vlaams Instituut voor Biotechnologie in Belgium have developed a machine learning model that can predict how good a beer will taste based on its chemical composition and offer suggestions for improving the recipe. The model, trained on data from 250 commercial Belgian beers and over 180,000 public reviews, can help brewers adjust their recipes, produce more consistent products, and save time and money on consumer trials.
5. Synthesizing Residents’ Feedback
CitizenLab, a U.S. based software services startup, has launched an AI-powered tool called the AI Sensemaking Assistant that can quickly scan, organize, and report on thousands of resident responses to community planning proposals. The tool aims to help public administrators get through large amounts of community input quicker, improving their decision-making process and minimizing their workload.
6. Discovering New Antibiotics
Researchers at Stanford Medicine and McMaster University in Canada have used generative AI to identify the chemical structures for six novel drugs aimed at killing resistant strains of Acinetobacter baumannii, a leading pathogen responsible for antibacterial resistance-related deaths. The AI model, named SyntheMol, generated around 25,000 possible antibiotics and laid out the chemical process needed to synthesize them in less than nine hours.
Researchers at the Indian Institute of Science (IISc) are using multiple swarms of drones to tackle forest fires. The drones are equipped with fire extinguishers, cameras, thermal sensors, and temperature detectors to spot fires and put out fires. Once a fire is discovered, the drones communicate with each other to calculate the fire’s size and potential spread, allocating the required number of drones to quench the fire while others search for additional fire clusters.
Google is integrating generative AI into travel planning, enabling users to get personalized travel and activity recommendations based on their requests. The AI system draws from reviews, photos, and other details to create tailored itineraries, thereby streamlining trip planning.
The New York City Police Department (NYPD) is testing AI-powered gun detection scanners in subways. The scanners, made by Massachusetts-based security technology firm Evolv, use advanced sensors to identify weapons on passengers entering the transit system. The deployment aims to improve safety and public confidence following several high-profile violent incidents.
10. Increasing Chatbot Efficiency
AI21 Labs, a startup based in Israel, released a new AI model that can remember and consider much more of the conversation history compared to other chatbots while running on a single GPU. The model achieves this result by combining two different AI architectures.
Image credits: Karsten Winegeart