With the race for artificial intelligence (AI) shifting into high gear, companies are not only developing AI systems in-house, but are increasingly bringing start-ups into the fold to augment their efforts. Market intelligence firm CB Insights released its list of AI acquisitions over the last five years. A review of the most active acquirers can provide insight into future developments at those companies, and show how the new technologies are being integrated.
One intriguing revelation is that many of the acquired companies work in the areas of deep learning and machine learning, two disciplines within AI. Algorithms and software are used to develop models of the human brain called neural networks, attempting to recreate our capacity to learn. Another revelation is that much of the acquired technology can be directly integrated into virtual personal assistants. This is an area of great interest to big tech, but is also being pursued by a surprising number of other companies.
Apple is giving Siri a makeover
Apple (NASDAQ:AAPL) has been at the leading edge of AI since its introduction of Siri five years ago. While it is making a significant number of acquisitions, each appears to specialize in a different area:
- VocalIQ concentrated its deep-learning capabilities to better understand human speech. This could provide Siri with an upgrade.
- Perceptio can run neural-network algorithms and perform advanced calculations locally on a cellphone, without uploading user data to the cloud. This would help Apple protect consumer data by keeping it local to each phone.
- Emotient attempts to read facial expressions in order to determine a person's emotional state. This recent acquisition is likely related to a patent Apple applied for in 2014, describing a system for determining mood based on numerous factors, including facial expressions.
- Tuplejump provided a better way to apply deep learning to large data sets as they are uploaded. Apple has previously stated that it could receive filtered data from iPhones, and this technology would enable it to process the data in real time.
One especially compelling acquisition was Turi, described as a "machine learning platform for developers and data scientists." It had created tools that allowed developers to build apps utilizing deep learning that could scale for many users; this tool kit was described as being similar to Google's TensorFlow open-source machine-learning library. Turi allows Apple to bring AI capabilities to developers for its App Store offerings.
Twitter wants a seat at the big tech table
Twitter (NYSE:TWTR) has made a number of AI-related acquisitions in the area of machine learning. Each appears to provide the company with specific opportunities to improve its current business:
- Madbits focused on technology related to searching and understanding images, and the ability to organize large data sets. It previously released an iOS app that created photo collages.
- Magic Pony developed neural networks that can sharpen blurry cellphone images or video to near HD-quality. It also can use information in the original to improvise new or expanded images -- great for use in virtual reality. These will both help Twitter increase its image features.
- TellApart allows marketers to better target potential customers across device platforms. This could improve monetization of Twitter's ads.
Whetlab is probably the most interesting acquisition. While it is less business-specific, it will bring the most advanced AI capability. It created technology that helps companies jump-start their deep learning by providing an infrastructure; the time required to train AI systems is reduced from months to days. This technology can be used to improve a variety of areas in Twitter's business, including language understanding, image recognition, and user profiles.
These acquisitions reveal how two very different technology companies are integrating AI with a goal of improving their respective businesses. Acquisitions for their own sake are folly, but using them to gain a competitive advantage, improve business processes, or to better serve customers can lead to success. Time will tell how successful each of these companies has been in leveraging these technologies.