Note: Content might be outdated!
Can someone explain to me what machine learning is, other than 'tracking everything you do on the internet so that we can advertise to you specifically'.
— Jonathan Bossenger (@jon_bossenger) July 11, 2019
From the angle of a web developer, it’s easy to think of machine learning exclusively as a means to more efficient advertising but it’s way more than just that. Besides ads and e-commerce, another big area of impact for machine learning on the web is journalism. Random examples from my bookmarks:
- Quartz, among other things they do around AI, offer a crash course for journalists in classifying text with machine learning.
- Thomson Reuters Open Calais helps structuring unstructured text with semantic metadata in RDF format.
- Machine learning can help tracing a circling helicopter (a traditional trigger for local journalism).
When you look beyond the web, machine learning is scoring big time all around, even in more ‘human’ sectors like agriculture, or health care. Again, these are just random examples from my bookmarks:
- High school students developed PlantMD, an app that detects diseases in plants using Google’s open source machine learning library, TensorFlow.
- FarmDrive uses mobile phones, alternative data, and machine learning to close the critical data gap that prevents financial institutions from lending to creditworthy smallholder farmers.
- Research from the Institute for Data, Systems, and Society aims to help African farmers increase their production and profits with better prediction.
- A Nigerian startup has developed a machine learning system to detect child birth asphyxia earlier and hopes to save thousands of babies’ lives every year. (Birth asphyxia is the third highest cause of under-five child deaths and is responsible for almost one million neonatal deaths annually, according to WHO.)
- The folks from Morphl busted five myths about AI and ML quite nicely.
- They also have a piece on the current state of web user interface design and development that really resonates with the UX person in me:
What if product metrics could be directly integrated in the product development process? What if we could add a new dimension to UI/UX development: METRICS. And what if we could build the UI/UX components to automatically adapt and morph based on these metrics?