Hey Reader!
As artificial intelligence and generative AI continue to advance rapidly, the technology sector is increasingly investing in Gen AI. Gartner predicted that by 2026, over 80% of organizations will be using Gen AI or Gen AI-enabled applications, compared to less than 5% in 2023.
The Oracle database is one of the world's most popular databases, known for keeping up with industry trends and providing innovative solutions to businesses. As the internet boom arrived, Oracle versions were named after it, such as 9i for "Internet" (my first experience working with a database started with 9i), 10g for "Grid computing," and 12c for "cloud computing." Now, with the world moving towards adopting artificial intelligence, the latest version, 23ai, is named "artificial intelligence."
This brings the required technological advancements to the core of the database, and you don't need to go out and look for them elsewhere.
Let's cover a few of the features:
AI Vector Search:
Large Language Models (LLMs) are transforming everyday tasks, but they have limitations, such as being able to answer questions based on information from when they were created and lacking specific details about data within an organization.
To overcome these challenges, we can enhance these models by integrating them with data relevant to your specific needs.
Oracle Database 23ai introduces AI Vector Search, a new feature that allows these AI models to generate and store vectors. Vectors, sometimes called embeddings, are multidimensional representations of various data types, such as documents, images, and videos. They help identify similarities between different data types through mathematical calculations.
The strength of Oracle Database 23ai lies in its ability to merge these similarity searches with traditional business data queries using simple SQL commands. This approach provides AI models with the extra context they need, enhancing their accuracy and making their responses more relevant. Oracle Database 23ai has made this possible by introducing a new data type, vector indexes, and extensions to the SQL language that simplify querying vectors along with regular business data. This functionality taps into Oracle's advanced analytical capabilities, making extracting meaningful insights from your data easier.
With the addition of AI Vector Search to Oracle Database, users can quickly and easily gain the benefits of artificial intelligence without sacrificing security, data integrity, or performance.
Example Use Cases:
- Conversational AI, or Chatbots: Build AI-powered digital assistants
- Similarity Search: Match customers with products
- Content-Based Filtering: Enable personalized recommendations, locate retail items from pictures
- Natural Language Processing: Text classification and clustering SQL generation
- Data Analysis: Anomaly detection, pattern recognition
- Computer Vision: Face recognition, biometric identification, object detection
- Biomedical Research: Gene/DNA similarity research, molecular structure search
- Geographic Information Systems: Spatial analysis, map rendering
- Industrial Applications: Quality control, predictive maintenance, machinery malfunction
Select AI:
Use natural language to query and analyze your data. You can easily embed natural language query capabilities into your applications. Oracle Database 23ai is integrating natural language capabilities with help from AI models like Cohere and Llama to make it easier for more people to use.
This means you can ask questions in simple, everyday language, like
"What are the top-selling items to college students this year?"
Oracle Database 23ai will interpret your question, translate it into a precise SQL query by categorizing "college students" within a specific demographic, and then run the query to give you the results.
Encoding of Data via ONNX standard
To leverage the benefits of large language models (LLMs), it's crucial to understand and process our datasets securely. Traditionally, the complexity of encoding data often led to passing this task to third-party services, risking sensitive information exposure.
Oracle Database 23ai addresses this concern by supporting the ONNX (Open Neural Network Exchange) standard, which allows users to upload their trusted AI models directly into the database. This capability ensures that data is encoded securely as entered into the database, eliminating the need to send data outside for processing. The ONNX standard is widely recognized for enabling interoperable machine learning models, supporting various machine learning tasks like classification and regression directly within the database environment. This approach enhances data security and speeds up the processing by enabling real-time data encoding and use within the database.
Machine Learning in Oracle Database:
Machine Learning in Oracle Database supports data exploration, preparation, and machine learning (ML) modeling at scale using SQL, R, Python, REST, automated machine learning (AutoML), and no-code interfaces. It includes more than 30 high-performance in-database algorithms that produce models for immediate application use.
By keeping data in the database, organizations can simplify their overall architecture and maintain data synchronization and security. This enables data scientists and other data professionals to build models quickly by simplifying and automating key elements of the machine learning lifecycle.
More than 300 features have been introduced in 23ai; detailed documentation is here.
