Computer simulation enters the era of AI: Altair’s story AI & Surroundings

Computer simulation enters the era of AI: Altair's story

Slides of a talk this year at the Altair technology conference by leading data scientists Mamdouh Refaat. The data that is normally collected in the real world to inform engineering options can be complemented with data created within a simulation.

Altair Engineering

“I don’t need the money, I wouldn’t be doing this if it wasn’t really intellectually interesting for me.”

Jim Scapa, founder and executive director of software manufacturer Altair Engineering, recently explained during an Italian food lunch why, at 62, he is working harder than he has worked in his life.

Scapa would certainly not be left alone to enjoy the slings and arrows of Wall Street investors. Although Altair’s shares, traded under the ticker “ALTR”, have doubled since the debut of the Nasdaq action on November 1, 2017, the stock has had some heartbreaking ups and downs. Investors have periodically worried that their ambitious mergers and acquisitions strategy to transform the company through artificial intelligence may be leading Altair astray.

Altair shares rose 32% this year, slightly less than the Nasdaq composite index.

“I had my training wheels as a public CEO, and it worked very well for the first six months or so,” Scapa told ZDNet during lunch in Manhattan. Things went very well right after the IPO, then the complaints came.

In short, the tension comes from the fact that most people are not sure what artificial intelligence is supposed to do for software in general terms. Scapa has a vision and requires some faith from investors.

Scapa, a mechanical training engineer, who also earned a master’s degree in business administration from the University of Michigan, has spent 35 years making Altair one of the preeminent companies in computer simulation. Its software dominates the design of Detroit’s chassis and car systems.

The common name for the field is “computer aided engineering” or CAE. This is the process of taking car models or other objects that have been written in a CAD / CAM tool and simulating how they will behave in the real world, things like the air flow around a vehicle as it moves through space , or the consistency of materials in an object constructed as that object ages.

Artificial intelligence is a new and emerging part of that. Machine learning, especially deep learning, can take huge amounts of data and use it to discover a function that is close to solving a problem. A designed object can live within a simulation on the computer that has a constant connection with reality. It’s more or less like Neo in The Matrix has a real life on earth, but also a life as an avatar inside the machine. It is the notion of a “digital twin,” and many have expressed it in the design software business for years as a kind of Holy Grail.

Co-founder and CEO of Altair, Jim Scapa.

Altair Engineering

The novelty of Scapa is that there will be “simulation-based digital twins and data-based digital twins,” as he says, forms of objects that take shape as the computer feeds on data and as new data They emerge from simulations.

The data allows machine learning to refine a design, somehow more objectively than a human designer could design one. Feeding the machine with simulation data, in a constant dialectic between design and simulation, can accelerate the process of refining designs.

Scapa believes that this dialectic of design and simulation is permeating the world thanks to artificial intelligence.

“We are at a time in time when, for many things we do, algorithms are driving decisions,” is the way in which Scapa broadly sees the landscape of society, not just engineering. “For me, it’s all about how algorithms help humans make decisions,” said Scapa.

To expand Altair’s reach in all aspects of the algorithms, Scapa has been on a shopping spree, buying thirty companies since he founded Altair in 1985. The immediate advantage of making Altair public after decades privately is that he filled out the Altair coffers. Cash and cash equivalents have increased to $ 247 million in the quarter of September from just over $ 17 million before the OPV. (Altair’s balance sheet was driven by an offer of follow-up actions last year).

Some of the agreements have been fine with investors, such as SimSolid, a company that Scapa bought in October last year after a six-month review process. You can perform simulations of designed objects without first dividing the computer model into simple parts. That can dramatically accelerate the simulation time. Scapa calls SimSolid “the product of a brilliant engineer from Belarus, a guy who had discovered a really difficult physics,” referring to SimSolid co-creator Victor Apanovitch, a former professor at the Polytechnic University of Belarus.

Other agreements have also not fallen, such as the acquisition in November 2018 of the DataWatch startup for $ 176 million in cash, the largest Altair agreement. To manage all the data that will be sent to machine learning systems, Altair needs to focus on data science. DataWatch, which began its life the same year as Altair, began as a manufacturer of IBM-compatible computer terminals, but then went on to manufacture only software. In particular, he spent years creating and buying business programs for so-called business intelligence, such as data reports, particularly for the financial services industry. DataWatch is a pioneer in data analysis and data science that, for Scapa, fits very well with Altair’s simulation capabilities.

Many did not see it that way. “The action fell right after we announced the agreement,” he recalled. Investors were uncomfortably surprised. They wanted Scapa to constantly build a very predictable simulation business. “They wanted me to keep buying solvers,” he said, using the industry’s vernacular language for computing engines that represent different physics problems and that can be continually added to the company’s platform in a predictable way to expand its usefulness. They didn’t want him to go on quixotic missions.

DataWatch was a company with a checkered past that some investors already distrusted. Richard Davis, a stock analyst at Canaccord Genuity who follows Altair’s actions, wrote last spring that “we have too much experience with DataWatch’s epic erroneous execution to believe that this purchase was anything but a mistake.”

“The lesson I learned is that I have to do a much better job sending messages to this, sometimes, so that people see where we are going,” Scapa told ZDNet. However, he seemed not to be discouraged and said: “We have some very good long-term investors.”

Some skeptics have come to reluctantly accept the method of Scapa. Davis, from Canaccord, wrote that despite his objections to the DataWatch agreement, “we give credit to Altair’s management for cleaning the house there and then assigning what appears to be a rational amount of employee talent to an effort to grow the business. ”

Davis compared Altair’s potential with Ansys, another publicly traded software manufacturer, whose shares multiplied by 11 in the first decade of the new century.

So far, mergers and acquisitions are adding to healthy growth rates. Analysts expect Altair to achieve revenue growth of 14% this year and perhaps 11% next year, when it can reach sales of $ 500 million. “I think it has the potential to be substantially larger and much broader,” says Scapa about the simulation market compared to what it has historically been.

With the ownership of approximately 45% of the outstanding shares of a $ 2.6 billion company, and more than half of the voting power, Scapa, who is also president, is in a position to do what he thinks he should do Despite the complaints. The price of an education in the stock market, in his opinion, was worth it in exchange for capital. “We were a rocket on the launch pad,” before the IPO infused the balance. Now it has rocket fuel to seek agreements, to build the portfolio.

As a nerd, Scapa cannot avoid marveling at the border of AI and the cloud and all that entails: partnering with Nvidia for the future of GPU computing; develop tools for NASA’s “Pleiades” supercomputer; Working with cloud computing giants to design a new silicon for AI: all these views have the feeling that Scapa and Altair have prospered enough to participate in an era of time.

“The future is bright,” Scapa said as lunch ended.


social experiment by Livio Acerbo #greengroundit #thisisnotapost #thisisart