🗄 ML to Power a New Generation of Databases
The Scope covers the most relevant ML papers, real-world ML use cases, cool tech releases, and $ in AI. Weekly.
The history of the software industry has been the history of data infrastructure. Each relevant technology trend in the last five decades has been accompanied by incremental progress in database technologies. Server-side software coincided with the emergence of relational databases; social and mobile technologies powered the NoSQL movement; cloud computing was the main catalyzer for the emergence of big data platforms. In the era of machine learning (ML), we are likely to see the evolution of a new type of database platforms optimized for data science workloads. However, ML has the unique capability to not only improve the existing generation of database technologies but also reimagine the space with new databases we haven’t seen before.
The influence that ML can have in database technologies is unique because it is bidirectional. Areas such as natural language processing (NLP) can power new query models for the existing database platform. Who doesn’t like the idea of interacting with data using natural language? However, the influence of ML in the database field can be more profound. Imagine embedding inference workloads as a native construct of a database engine. Google BigQuery ML is a great example of this type of concept. Taking these ideas further, ML pipelines themselves can benefit from a new type of database that is more optimized for training and evaluation workflows. Research in all these areas is accelerating at a frantic pace. Just this week, Facebook published a research paper unveiling what they call neural databases, a concept that combines the use of NLP for unstructured databases. Like previous technology trends, ML is likely to bring fresh ideas that power innovation in the world of databases.
🔺🔻TheSequence Scope – our Sunday edition with the industry’s development overview – is free. To receive high-quality content about the most relevant developments in the ML world every Tuesday and Thursday, please subscribe to TheSequence Edge 🔺🔻
🗓 Next week in TheSequence Edge:
Edge#119: we discuss the topic “Data Labeling - Build vs. Buy vs. Customize”; we explore how by identifying behaviors in previously labeled data we can build a pipeline to label the rest of the data; we overview Label Studio.
Edge#120: we go practical and talk through several use cases of data labeling customization.
Now, let’s review the most important developments in the AI industry this week
🔎 ML Research
Facebook AI Research (FAIR) published a paper proposing Neural Databases, a new concept to search unstructured databases using natural language ->read more on FAIR team blog
Amazon Research published two papers detailing case studies for PECOS, their open-source framework for multilabel ranking ->read more on Amazon Research blog
🛠 Real World ML
The Waymo team published some details about Waymo Driver, their autonomous driving system that just expanded its testing in San Francisco ->read more on Waymo blog
ML at Headspace
Meditation startup Headspace published some details about their real-time ML infrastructure ->read more on their blog
Using Data in Uber’s Rider App
The Uber engineering team published a blog post detailing the data processing infrastructure powering their Rider app >read more on their blog
🤖 Cool AI Tech Releases
IBM unveiled Telum, a new processor for large-scale deep learning inference ->read more on IBM Research blog
NVIDIA AI Enterprise
NVIDIA announced the general availability of its AI Enterprise platform that enables running a new set of tools and frameworks in the VMWare vSphere platform ->read more in NVIDIA press release
💎 We recommend*
A limited-time offer: as a reader of TheSequence, you can receive a consultation with a Squark AI Expert to review your use cases and potential predictive results at no charge. You will also receive a 15% discount on a Squark license. This is the only discount available, and it’s exclusively for TheSequence audience. Check with an expert how your company can benefit from Automated Predictive Analytics:
💸 Money in AI