The Partnership Between ESI and ENSAM

The global simulation company ESI Group recently announced, through a collaboration with ENSAM (Ecole Nationale Supérieure d’Arts et Métiers), that its INENDI Inspector software is released as an open-source tool. The software is available through a github repository. The French company, which is committed to supplying the industry with tools to make better decisions through simulation and data analysis, previously worked with ENSAM on a smart city/smart nation project in Singapore.

INENDI Inspector creates several visualization options. (Image courtesy of ESI.)

In the wildly crowded field of data analytical software, INENDI Inspector hopes to distinguish itself as a data visualization tool right alongside the machine learning functions. Open source is a popular release method for machine learning and artificial intelligence (AI) software, and ESI is entering the field with a university partner and a complementary tool. Big data and AI are two huge behemoths rolling through industries right now, and the company hopes to make a difference in each arena with applications big and small. Clearly, one application on the radar is simulation.

What Is Inspector and What Does It Do?

Inspector brands itself as a tool that “allows any user to perform interactive explorations over very large amounts of data.” This is almost boilerplate wording for an AI software, but Inspector is released with equal focus on its role as a data tool and a machine learning tool. MINESET, the previous iteration of this software, touted itself as data visualization without the need for coding, but the learning curve seemed fairly steep.

Large sets of data can be input and then visualized based on a few sets of menu options, with the first step being data inspection to find and remove outliers. Visualizations show users hidden relationships between variables, and several options exist to show different output styles. Users can find weak signals in the data by dialing competing datasets up or down. Machine learning is also an important part of the tool, as algorithms are built and controlled in Inspector.

INENDI Inspector is coding heavy, and the latest 4.9 release of the software boasts a Python-wrangling feature that allows users to open Python scripts after data intake occurs. The system is compatible with Python modules, including NumPy, Pandas, TensorFlow and Scikit-learn. And now the software is open source and downloadable through github.

The company says the software has strong use cases in the areas of “predictive maintenance, cybersecurity, control, and analysis of machine learning algorithms.” Cybersecurity is a big focus when dealing with risk control and weak signals. The company’s Australian branch discusses some case studies where risk control, weak signals and advanced packaging tools are explored to minimize risk for customers. Using data visualization tools to understand IT logs can help customers to see big picture things like network usage trends and peak stress times. But anomalies like weak signal attacks and accounting fraud can come to light when the data is further explored.

Why Move to Open Source?

At this stage of AI development, releasing your tools as open source is common. Even though regulations and governance are still somewhat fuzzy, the overall feeling is that AI is a democratization of technology. Many major computing players—Google, Microsoft, Amazon and IBM—all have some tools available for free use. Or rather, for use after registering with the system. Open source is necessary for software to be visible and gain wide use in the AI marketplace.

Gaining customers and entering the conversation as a popular tool are both desirable outcomes from moving to open source. The drawback to an open-source release is the loss of revenue from business licenses. Releasing a two-tiered software is generally the way that a company keeps control and retains its revenue, but ESI has pushed Inspector completely to github.

ESI’s announcement points out that the open source is a request that customers have made, and the company also has a strong commitment to updating and evolving its software based on user requests and feedback. The university partnership is also a strong push to ensure that Inspector is open source. University budgeting is notorious for having a long lead, and knowing that you’ve got an educational tool with a perpetually zero line-item price is a great selling point. When a generation of university students use the software to learn data visualization and machine learning, then it stands to reason that these young professionals might stick with INENDI Inspector out in the real world.

Open source is also an opportunity for the company’s user base to get inside the guts of the programming and look around, and this creates the possibility that improvements or enhancements will be found by the world at large. The software remains relevant and will evolve with the needs of users, and the information that users provide will help the company ensure that the software’s direction aligns with its future planning.

Why Simulation Users Should Care

The takeaway from the arrival of a new tool for data mining and machine learning is twofold. First, there is an incredibly busy marketplace for data tools right now. AI is gaining traction in manufacturing and business but is also being used to help solve bigger picture problems. The Engineering Grand Challenges and the United Nations Sustainable Development Goals will all lurk in the background for society over the next 50 years. It is encouraging to see a simulation company partner with a university to build data instruments that can help solve problems both big and small.

The likely other part of the puzzle for ESI group is the integration of this data visualization and machine learning tool with the rest of the company’s simulation tools. One of the company’s missions is “the vision of sustainable product innovation and manufacturing processes with zero physical tests and prototypes.” There are many easy and immediate parts of ESI’s product portfolio where this tool can be integrated to create higher-quality simulations. The company’s largest customer bases are in automotive, aerospace, defense and heavy machinery. The development cycles of all these industries can benefit from gaining insight into data and machine learning.

Artificial intelligence continues to grow as an industry, and even though an ISO standards committee exists, regulation is still not fully formed. Companies that understand the benefits and risks of AI and machine learning position themselves well in the upcoming data-rich future. If there is ever a data-rich application that ESI pays attention to, it is simulation, so AI is a good fit. Releasing INENDI Inspector as open source should help ESI and ENSAM to gain insight into both machine learning and data visualization in the near future.