pwshub.com

Google debuts new accuracy-optimized DataGemma LLM series

Google LLC has developed a series of language models that can answer questions about numerical facts more accurately than earlier algorithms.

The DataGemma series, as the model lineup is called, debuted on Thursday. Google has made the source code for the algorithms available on Hugging Face

DataGemma is optimized to field user questions about statistical facts such the average revenue of companies in a given market segment. It answers queries using information from Data Commons, a free knowledge repository maintained by Google. The repository contains more than 240 billion data points from sources such as the United Nations, the World Health Organization, the CDC and census bureaus.

Under the hood, the DataGemma series is based on Gemma 2 27B, an open-source large language model that Google released in June. Gemma 2 27B powered by the industry-standard Transformer neural network architecture and features 27 billion parameters. Google says that it can rival the performance of LLMs with twice as many parameters. 

According to the company, DataGemma is based on a version of Gemma 2 27B that was specifically optimized to process numerical facts. The model interacts with Data Commons, the knowledge repository from which it retrieves those facts, through a natural language search bar. 

“Instead of needing knowledge of the specific data schema or API of the underlying datasets, DataGemma utilizes the natural language interface of Data Commons to ask questions,” Google software engineer Jennifer Chen and Prem Ramaswami, the head of Data Commons, detailed in a blog post. “The nuance is in training the LLM to know when to ask.”

Google developed two versions of DataGemma as part of the project. Each takes a different approach to answering user questions.

The first version leverages a method known as RIG, or retrieval-interleaved generation, to process queries. When a user asks a question, the model doesn’t generate an answer based on its internal knowledge base but rather asks Data Commons for the needed information. The LLM then uses the retrieved data to generate a prompt response. 

The second version of DataGemma implements the more widely-used RAG, or retrieval-augmented generation, data management method. When a user enters a query, the model retrieves information relevant to the prompt from Data Commons. It then sends the collected information to Google’s more advanced, proprietary Gemini 1.5 Pro model, which generates an answer. 

According to MIT Technology Review, the RIG version of DataGemma can correctly retrieve numerical facts from Data Commons about 58% of the time. That’s compared with the 5% to 17% accuracy achieved by the other models that Google evaluated. The RAG version of DataGemma, in turn, generated correct answers to 80% to 94% of the answers it received during the search giant’s tests.

Google plans to improve DataGemma by training it on more information. Additionally, Google will increase the number of questions the LLM series can answer from a few hundred to millions. Further down the line, the company plans to integrate DataGemma’s data retrieval capabilities into its flagship Gemini series of language models.

Source: siliconangle.com

Related stories
1 month ago - A flurry of new artificial intelligence models this week illustrated what’s coming next in AI: smaller language models targeted at vertical industries and functions. Both Nvidia and Microsoft debuted smaller large language models too....
1 month ago - Regulators are circling ever closer to big tech companies — the latest being Google, which the Federal Trade Commission more than hinted this week should be broken up. It’s not at all certain that will happen, since it’s up to the judge...
2 days ago - Artificial intelligence startup and MIT spinoff Liquid AI Inc. today launched its first set of generative AI models, and they’re notably different from competing models because they’re built on a fundamentally new architecture. The new...
2 weeks ago - This was the week that Apple finally infused artificial intelligence into its new iPhones, Watches and AirPods, though some of features won’t be coming for a bit and overall, the AI stuff seemed a little underwhelming. The medical...
6 days ago - This week brought yet another big shakeup at OpenAI, as Chief Technology Officer Mira Murati and others quit. But CEO Sam Altman seems to be cementing his control. And Chief Financial Officer Sarah Friar said in a memo that OpenAI’s...
Other stories
18 minutes ago - Google LLC will enable brands to display ads in AI Overviews, the natural language explainers that its search engine generates in response to some queries. The change is part of a broader product update that the Alphabet Inc. detailed...
18 minutes ago - Artificial intelligence is a game-changing technology in the enterprise world, increasingly adopted for its potential to enhance human capabilities. AI trends are driving industrial transformation across sectors. Large language models,...
18 minutes ago - U.S. Federal Reserve losses crossed the $200 billion point this week, according to data released on Thursday by the central bank. The Fed reported that as of Wednesday, the level of its so-called earnings remittance to the Treasury...
1 hour ago - At last week’s Climate Week and the Global Citizen Festival in New York, both committed to making the world a better and more livable place for future generations, technology was viewed by many as the driving force to shift to a cleaner,...
1 hour ago - Employees can spend up to one-third of their time looking up information in the company, and until now enterprise search has not been up to snuff. That looks to change with Glean Technologies Inc.’s artificial intelligence-based...