In December, AI made major waves: It became poised to reach more people than ever. In 2019, we’re certain AI is going to make a significant impact on businesses and communities alike.
1. AI in 2019: Trends and Expectations
Analytics Insights’ article on the top AI trends and expectations for 2019 covered everything from virtual agents to cybersecurity.
Speech recognition, image recognition and smart recommendations are also mentioned, in addition to Forrester’s top five predictions:
- Investment in information architecture (IA) as companies realize the importance of supplying AI with high-quality data.
- Demand for AI talent will grow as an increasing number of organizations struggle to find candidates capable of training and managing AI systems.
- The digital workforce will proliferate as robotic process automation (RPA) and AI are combined to create an ultra-efficient labor force.
- Explicable AI will be in high demand in response to data regulations as well as a practical need for transparent, easily understandable models.
- Human wisdom will prevail as AI leaders realize the importance of expertise and input from real people.
Also be sure to check out this article at MarTech Advisor for more insights into the future of AI.
There, five industry influencers, including SambaNova’s own Marshall Choy, share how they think AI will impact businesses in 2019.
2. ML Accurately Detects and Predicts Earthquakes
Machine learning is good for more than tech work: It also has the power to save lives.
We’ve covered developments with earthquakes and AI before, here, but now researchers from the Los Alamos National Laboratory revealed in December that they’d successfully trained an ML model to detect seismic tremors in the Cascadia region.
Although those tremors had previously been dismissed as benign, the machine learning model recognized that they contained a distinct pattern that indicates both slippage and fault failure.
With this powerful machine learning-powered tool at their disposal, scientists may now be able to accurately predict Cascadia’s impending megaquake, potentially saving thousands of lives in the process.
3. Great Data Engineering is Critical to Great Machine Learning
Machine learning isn’t a new idea, but the massive amounts of data used to support it has only recently become available.
This presents enterprises and adopters with a myriad of possibilities, as well as a significant challenge: In order to develop a great ML model, you need excellent data engineering to back it up.
An article outlining a new TDWI Checklist Report highlighted the insights of Fern Halper, vice president and senior director of TDWI Research for advanced analytics:
The advent of big data has, in several important ways, both revitalized machine learning and increased the complexity of using these models to drive insight and action. Data engineers must create robust production data pipelines to feed machine learning models the increasing amounts of disparate data they require.
Going forward, organizations need to place as much of an emphasis on data engineering best practices as they do on ML itself.
4. Neural Networks Learn to See in the Dark
AI is making an impact in countless industries, but its role in the future of medicine is arguably one of the most exciting.
At the Massachusetts Institute of Technology (MIT), researchers have trained a neural network to see in the dark by reconstructing objects from near-pitch black images.
The result was neural network-generated objects that were “more defined than a physics-informed reconstruction of the same pattern, imaged in light that was more than 1,000 times brighter.”
Since patients and biological samples can be adversely affected by traditional imaging methods, the neural network’s ability to clearly identify and reconstruct objects from exceedingly dark images has great potential in the medical field.
5. Seamless Human-AI Relationships Still Eludes Us
Even though AI technology has made huge strides in the last year alone, there’s still a long way to go before humans can form trusting relationships with AI.
Elite officials and scientists are struggling not only to create an AI program that can intuitively interact with humans but also to identify the steps they can take to get there.
As Arizona State University associate professor Spring Berman pointed out, “explicable AI is a big challenge … it’s like a black box.” Plus, we have to learn much more about how humans learn and work before we can teach AI to do the same.
Although we’re far from discovering all that AI can do and learn, AI has come a long way. From earthquake prediction to medical imaging, AI has performed a number of impressive feats in 2018, and will undoubtedly continue to do so in 2019.