5 Frequently Asked Questions about Altair’s AI-Powered Engineering Solutions
Altair Employee
1: Can the AI for engineering applications run on my laptop, or do I need cloud access?
AI-powered engineering solutions can run on a laptop, though for intensive tasks such as computational fluid dynamics (CFD), using high-performance computing (HPC) in the cloud is more efficient. But yes, Altair's physicsAI, romAI, the HyperWorks platform, or the RapidMiner platform can be accessed on your laptop. These platforms include easy-to-use workflows for cloud computing access when needed.
2: Can we select a specific model for predicting new simulation results, given that both the test set and the simulation set are used to train the AI model?
Once your AI model has been trained with a given data set, you can't distinguish which data type is used to make predictions. In other words, you can't request that predictions be based only on field data if test data was also used to train your AI model. However, you can create multiple training models based on the data source. For example, test data can be used to train an AI model coming from simulation and train another AI model coming from field data. And depending on how you want to create a AI-trained model, you can always provide all data types.
3: What problem do I use Altair romAI and Altair physicsAI for?
It depends on the problem, results, and insights you're looking for. If you are working on very complex and large systems, and you want to mimic that system very closely and accurately, then it’s possible romAI. romAI can efficiently generate accurate reduced order models using small data sets to simulate any kind of scenarios, predicting with multiple physics, for multiple applications.
With physicsAI you can predict one kind of physics at a time, generating field predictions to evaluate more designs rapidly during the design stage. So given a new model, you can quickly generate surfaces contours, stresses, anything to do with the structure’s results.
At Altair we have a clear path to choosing the right tools for the right kind of application.
4: As a design engineer, should I begin by learning AI/ML and then focus on applying AI to engineering tasks?
Engineers without data science experience using the Hyperworks platform with pre-processing, post-processing, and embedded AI can use existing workflows to access data, train AI models, and start predicting with only a few hours of training from Altair experts to understand and know which methods to use where. However, if you want to be an AI champion in your organization, additional training and experience are needed.
5: Does HyperWorks specialize in simulation while RapidMiner focuses on AI and machine learning?
Yes. RapidMiner is our machine learning and AI platform for data analytics. HyperWorks is our design and simulation platform that is powered by AI. For example, physicsAI, expert AI, and shapeAI technologies are embedded in HyperMesh’s design and simulation workflows without needing direct access to RapidMiner. In some cases, HyperWorks used in combination with RapidMiner works best. But you can do what you need on design, predictive analytics, field predictions in HyperWorks.
In summary, organizations that harness data to build strategic digital assets can fuel radical discoveries and continuous improvement. I address this challenge in-depth in this webinar: The Fish Are Byting: How to Reel in the Right Data for Your Digital Transformation.