![]() IBM’s recent acquisition of Ahana, a Silicon Valley startup that developed a cloud service based on Presto, will likely play here in the near future. That includes IBM’s well-known Db2 and Netezza engines, but also popular open source options like Presto and Apache Spark. Watsonx.data will be available both on-prem and in the cloud, and will enable users to bring the analytic engines of their choice to bear on data stored there. Watsonx.data serves as a lakehouse, which is a relatively new data architecture that blends elements of well-governed data warehouses built on trusted relational database technologies like Db2, Oracle, or Teradata along with the more scalable but messier data lakes based on HDFS or object storage systems, like Amazon S3 or Cleversafe, an S3-compatible object storage system that IBM acquired several years ago and which today forms the basis for IBM Cloud storage. It’s also planning to include Watson Code Assistant, Watson Assistant, Watson Orchestrate, and AIOps Insights as foundational models in wastonx.ai, it says. ![]() IBM is also planning on including thousands of open source models developed by Hugging Face, which creates generative models suitable for building chatbots and interactive AI applications. Lastly, the fm.geospatial model will be trained on climate data for assistance in planning for natural disasters. The fm.NLP collection will include LLMs for specific domains that can be tuned by users. The fm.code models will include code generation models that will be useful for developers who desire a “copilot” experience when writing code. ![]() Three groups of models will be included in the watsonx.ai studio. Users will gain access to LLMs as well as models suitable for training upon source code, time-series data, tabular data, geospatial data, and IT events data, IBM says. The studio ships a collection of “foundation models” that have already been trained on a “large, curated set of enterprise data backed by a robust filtering and cleansing process and auditable data lineage,” IBM says. Watsonx.ai provides a place for developers to train, test, tune, and deploy traditional machine learning models (think logistical regression and K-means clustering) as well as new generative AI capabilities. The first two watsonx components are expected to be available in July, while the last one is due in October. Watsonx is composed of three parts, including a development studio called watsonx.ai, a data lakehouse called watsonx.data, and a governance toolkit called (you guessed it) ernance. It’s not even the first time Big Blue has used the name (anybody remember IBM Watson Analytics?) But the new watsonx offerings represent IBM’s attempt to build off the renewed interest in AI, including but not limited to generative AI like ChatGPT. Among the firms looking to find a foothold in this fast-moving world is IBM. Now the rush is on for businesses to take advantage of LLMs as well as their image-generating AI cousins, such as OpenAI’s DALL-E 2 and Stable Diffusion. While LLMs like ChatGPT aren’t “intelligent” in the classic sense, they can still do some pretty impressive things, such as compose sonnets in the style of William Shakespeare, write high school term papers, or even generate syntactically correct RPG code. Thanks to the public launch of ChatGPT in late last year, regular citizens (as opposed to AI experts and industry analyst) have become aware of the vast potential of large language models (LLMs) that can mimic humans in very convincing ways. You don’t need an industry analyst to tell you that artificial intelligence (AI) is having a moment in the sun this year. ![]() IBM Introduces watsonx for Governed Analytics, AIĪt its annual Think conference last week, IBM took the wraps off watsonx, a new platform for crunching big data and developing AI and machine learning applications in a safe and governed environment.
0 Comments
Leave a Reply. |