Artificial intelligence (AI) has come a long way in recent years, and the advancements show no signs of slowing. Here are five recent examples of progress in AI that could positively impact the future.
1. New System for Self-Driving Cars Recognizes Critical Traffic Situations in Advance
Although many people are curious and excited about self-driving cars, they want to ensure they operate safely. Today’s models employ numerous AI features that help vehicles recognize their position on the road, notice other cars and more.
German researchers recently developed an AI-based system that allows self-driving vehicles to detect critical situations that require human intervention to resolve. The main characteristic that makes this achievement particularly remarkable is that it can recognize those events up to seven seconds before they occur.
The recurrent neural network-based system works by recognizing patterns in data. It also records types of situations that the vehicle was unable to handle independently before, then looks at incoming information for signals that could indicate similar conditions about to happen soon.
2. Smart Speaker Skill Analyzes Heart Rhythm Without Contact
Smart speakers demonstrated how AI could become a part of households around the world and make lives easier. Besides offering more convenience, they could help people stay healthier. Some skills let people find the nearest urgent care clinic or reorder prescriptions.
Thanks to work at the University of Washington, people may soon use smart speakers to become aware of irregular heart rhythms. Researchers created algorithms that confirm whether a person has a regular or irregular heartbeat without using any physical accessories.
Instead, a smart speaker emits inaudible sounds and can detect the heartbeat signals based on how they get reflected back to the device. This is not a commercially available smart speaker skill, but scientists said it succeeded when cardiac patients sat only 1 to 2 feet away from the gadget. The researchers also hope a future version could offer continuous monitoring as people sleep.
3. The Technology’s Evolution Minimizes Bias
People often assert that an AI algorithm is only as good as the humans that develop it. People aren’t perfect, and that means bias can unintentionally appear. However, as AI specialists have realized that possibility, they’ve taken proactive steps to mitigate it. For example, some companies take care to develop models that substantially reduce discrimination while giving accurate results.
Another recent example of applying AI to minimize bias involved looking at historical data associated with India’s Bollywood film industry. The project concerned examining 1,400 movies from the past 70 years and searching for gender and social prejudices within the subtitles.
The algorithms also assessed specifics like how often males appeared in films and what kind of skin beautiful women had. This efficient system of evaluating bias through the ages could help people in all societal sectors avoid repeating those mistakes.
4. AI Enables Quicker Vaccine Development for COVID-19 Mutations
Vaccine rollouts are underway, but many scientists believe people will need periodic boosters to keep them protected as the virus mutates. AI may prove crucial for getting those shots on the market faster.
Researchers in Southern California built an AI framework that eliminates 95% of the options that could target COVID-19, leaving only those likely to give the best results. It can also assess the approximate effectiveness of more than 700,000 proteins in a single group of data.
This system helps develop vaccines for the top three COVID-19 mutations. However, researchers said they could adjust it to cater to additional ones, too.
5. AI Helps Curb Misinformation Regarding Breaking News
The rise of internet news made misinformation rampant. Social media sites now use tags to flag unconfirmed or disputed facts, usually with help from AI. Newly published research suggests that strategy may be most effective when used with breaking news stories about which people do not have existing opinions.
The team looked at coronavirus-related news in their study. They determined it’s crucial to reach people early and explain why the AI flagged content as misinformation or fake news. Building a tool that detects misleading information is only half the battle. The other essential part is getting people to accept how AI categorized it.
Dorit Nevo, an associate professor involved with the study, said, “Our work with coronavirus news shows that these findings have real-life implications for practitioners. If you want to stop fake news, start right away with messaging that is reasoned and direct. Don’t wait for opinions to form.”
An Exciting Year for AI
These five AI feats indicate it’ll be a promising year for AI. People interested in the subject should stay abreast of happenings to see what other notable events occur.