While large language models (LLMs) and generative AI have dominated enterprise AI conversations over the past yr, there are other ways businesses can profit from AI.
An alternative are large quantitative models (LQMs). These models are trained to optimize specific objectives and parameters relevant to the industry or application, reminiscent of material properties or financial risk metrics. This is in contrast to the more general language comprehension and language generation tasks of LLMs. One of the leading proponents and industrial providers of LQMs is SandboxAQwhich announced today that it has raised $300 million in a brand new funding round. The company was originally a part of Alphabet and was spun off as a separate company in 2022.
The funding is a testament to the corporate's success and, more importantly, its future growth prospects in solving enterprise AI use cases. SandboxAQ has established partnerships with major consulting firms reminiscent of Accenture, Deloitte and EY to distribute its enterprise solutions. The predominant advantage of LQMs is their ability to deal with complex, domain-specific problems in industries where underlying physics and quantitative relationships are crucial.
“It’s all about core product development at the businesses that use our AI,” Jack Hidary, CEO of SandboxAQ, told VentureBeat. “So if you must develop a drug, a diagnostic, a brand new material, or do risk management at a big bank, then quantitative models are for you.”
Why LQMs are vital for enterprise AI
LQMs have different goals and performance in another way than LLMs. Unlike LLMs, which process text data sourced from the Internet, LQMs generate their very own data from mathematical equations and physical principles. The aim is to deal with quantitative challenges that an organization might face.
“We generate data and procure data from quantitative sources,” Hidary explained.
This approach enables breakthroughs in areas where traditional methods have faltered. In battery development, for instance, where lithium-ion technology has dominated for 45 years, LQMs can simulate thousands and thousands of possible chemical combos without the necessity to create physical prototypes.
Similarly, in pharmaceutical development, where traditional approaches face high failure rates in clinical trials, LQMs can analyze molecular structures and interactions on the electron level. In the financial services sector, nonetheless, LQMs eliminate the constraints of traditional modeling approaches.
“Monte Carlo simulation is not any longer sufficient to handle the complexity of structured instruments,” said Hidary.
A Monte Carlo simulation is a classic type of computing algorithm that uses random sampling to acquire results. The SandboxAQ LQM approach allows a financial services company to scale in ways in which Monte Carlo simulation cannot. Hidary noted that some financial portfolios might be extremely complex with all forms of structured instruments and options.
“If I actually have a portfolio and I would like to know what the tail risk is for changes in that portfolio,” Hidary said. “I would like to create 300 to 500 million versions of this portfolio with minor changes to it after which have a look at tail risk.”
How SandboxAQ uses LQMs to enhance cybersecurity
Sandbox AQ's LQM technology focuses on enabling corporations to develop latest products, materials and solutions moderately than simply optimizing existing processes.
One of the areas wherein the corporate has innovated is cybersecurity. In 2023, the corporate first launched its sandwich cryptography management technology. This has since been expanded further with the corporate's AQtive Guard enterprise solution.
The software can analyze a corporation's files, applications, and network traffic to discover the encryption algorithms used. This includes detecting using outdated or faulty encryption algorithms reminiscent of MD5 and SHA-1. SandboxAQ feeds this information right into a management model that may alert the Chief Information Security Officer (CISO) and compliance teams to potential vulnerabilities.
While an LLM might be used for a similar purpose, the LQM offers a unique approach. LLMs are trained on large, unstructured web data, which might contain details about encryption algorithms and vulnerabilities. In contrast, Sandbox AQ's LQMs are created based on targeted, quantitative data about encryption algorithms, their properties, and known vulnerabilities. The LQMs use this structured data to construct models and knowledge graphs specifically for encryption evaluation, moderately than counting on general language understanding.
Looking forward, Sandbox AQ can also be working on a future remediation module that may mechanically suggest and implement updates to the encryption in use.
Quantum dimensions without quantum computers or transformers
The original idea behind SandboxAQ was to mix AI techniques with quantum computing.
Hidary and his team recognized early on that real quantum computers wouldn’t be easy to acquire within the short term and wouldn’t be powerful enough. SandboxAQ leverages quantum principles implemented through improved GPU infrastructure. Through a partnership, SandboxAQ has expanded Nvidia's CUDA capabilities to incorporate quantum technologies.
SandboxAQ also doesn’t use transformers, which form the premise of virtually all LLMs.
“The models we train are neural network models and knowledge graphs, but they are usually not transformers,” Hidary said. “You can generate from equations, but you may as well get quantitative data from sensors or other kinds of sources and networks.”
Although LQMs are different from LLMs, Hidary doesn’t see it as an either/or situation for corporations.
“Use LLMs for what they’re good at after which herald LQMs for what they’re good at,” he said.
