Secure Enclave Service (SES) outage due to maintenance on March 23

By | News, Systems and Services, Uncategorized

Beginning at 6:30 a.m. on Saturday, March 23, through 5 p.m. on March 23, the Secure Enclave Services cluster will be shut down to allow for emergency maintenance on the underlying power distribution unit in the data center. This includes shutdowns for virtual machines hosted on the cluster.  To perform the upgrade, the entire cluster and all virtual machines will be offline on March 23.  

Status updates will be posted on the ITS Service Status page

How can we help you?

For assistance or questions, please contact ARC at arc-support@umich.edu.

For other topics, contact the ITS Service Center:

Using GenAI to design floor plans and buildings

By | Data, Data sets, Educational, Feature, General Interest, HPC, News, Research, Systems and Services

There is a lot to consider when designing places where humans live and work. How will the space be used? Who’s using the space? What are budget considerations? It is painstaking and time consuming to develop all of those details into something usable. 

What if Generative AI (GenAI) could help? We already know that it can be used to create text, music, and images. Did you know that it can also create building designs and floor plans? 

Dr. Matias del Campo, associate professor of architecture in the Taubman College for Architecture and Urban Planning, has been working to make architectural generative models more robust. He aims to expand on the patterns, structures, and features from the available input data to create architectural works. Himself a registered architect, designer, and educator, del Campo conducts research on advanced design methods in architecture using artificial intelligence techniques.

He leverages something called neural networks for two projects: 

  • Common House: A project that focuses on floor plan analysis and generation.
  • Model Mine: A large-scale, 3D model housing database for architecture design using Graph Convolutional Neural Networks and 3D Generative Adversarial Networks.

This is an example from the annotated data created from the Common House research project. The main obstacle that has emerged in creating more real-life plans is the lack of databases that are tailored for these architecture applications. The Common House project aims at creating a large-scale dataset for plans with semantic information. Precisely, our data creation pipeline consists of annotating different components of a floor plan, for e.g., Dining Room, Kitchen, Bed Room, etc.

 

Four quadrants showing 9 models each of chairs, laptops, benches, and airplanes

A large scale 3D model housing database for Architecture design using Graph Convolutional Neural Networks and 3D Generative Adversarial Networks.

What exactly are neural networks? The name itself takes inspiration from the human brain and the way that biological neurons signal to one another. In the GenAI world, neural networks are a subset of machine learning and are at the heart of deep learning algorithms. This image of AI hierarchy may be helpful in understanding how they are connected.

Dr. del Campo’s research uses GenAI for every step of the design process including 2D models for things like floors and exteriors, and 3D models for shapes of the rooms, buildings, and volume of the room. The analysis informs design decisions. 

DEI considerations

del Campo notes that there are some DEI implications for the tools he’s developing. “One of the observations that brought us to develop the ‘Common House’ (Plangenerator) project is that the existing apartment and house plan datasets are heavily biased towards European and U.S. housing. They do not contain plans from other regions of the world; thus, most cultures are underrepresented.” 

To counterbalance that, del Campo and his team made a global data collection effort, collecting plans and having them labeled by local architects and architecture students. “This not only ensured a more diverse dataset but also increased the quality of the semantic information in the dataset.”

How technology supports del Campo’s work

A number of services from Information Technology & Services are used in these projects, including: Google at U-M collaboration tools, GenAI, Amazon Web Services at U-M (AWS), and GitHub at U-M

Also from ITS, the Advanced Research Computing (ARC) team provides support to del Campo’s work. 

“We requested allocations from the U-M Research Computing Package for high-performance computing (HPC) services in order to train two models. One focuses on the ‘Common House’ plan generator, and the other focuses on the ‘Model Mine’ dataset to create 3D models based,” said del Campo. 

Additionally, they used HPC allocations from the UMRPC in the creation of a large-scale artwork called MOSAIK which consists of over 20,000 AI-generated images, organized in a color gradient. 

A large scale 3D model housing database for Architecture design using Graph Convolutional Neural Networks and 3D Generative Adversarial Networks.

“We used HPC to run the algorithm that organized the images. Due to the necessary high resolution of the image, this was only possible using HPC.”

“Dr. del Campo’s work is really novel, and it is different from the type of research that is usually processed on Great Lakes. I am impressed by the creative ways Dr. del Campo is applying ITS resources in a way that we did not think was possible,” said Brock Palen, director of the ITS Advanced Research Computing. 

Related: Learn about The Architecture + Artificial Intelligence Laboratory (AR2IL)

Using natural language processing to improve everyday life

By | Data, Great Lakes, HPC, News, Research, Uncategorized

Joyce Y. Chai, professor of electrical engineering and computer science, College of Engineering, and colleagues have been seeking answers to complex questions using natural language processing and machine learning that may improve everyday life.

Some of the algorithms that they develop in their work are meant for tasks that machines may have little to no prior knowledge of. For example, to guide human users to gain a particular skill (e.g., building a special apparatus or even, “Tell me how to bake a cake”). A set of instructions based on the observation of what the user is doing, e.g., to correct mistakes or provide the next step, would be generated by Generative AI, or GenAI. The better the data and engineering behind the AI, the more useful the instructions will be. 

“To enable machines to quickly learn and adapt to a new task, developers may give a few examples of recipe steps with both language instructions and video demonstrations. Machines can then (hopefully) guide users through the task by recognizing the right steps and generating relevant instructions using GenAI,” said Chai.

What are AI, machine learning, deep learning, and natural language processing?

It might help to take a step back to understand AI, machine learning (ML), and deep learning at a high level. Both ML and deep learning are subsets of AI, as seen in the figure. Some natural language processing (NLP) tasks fall within the realm of deep learning. They all work together and build off of each other.

Artificial Intelligence, or AI, is a branch of computer science that attempts to simulate human intelligence with computers. It involves creating systems to perform tasks that usually need human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

“NLP is a sub-area in AI, where DL/ML approaches are predominantly applied,” stated Chai.

Christopher Brown, a research data scientist, with ITS Advanced Research Computing (ARC) and a member of the ARC consulting team, explains that ML is a subfield of AI. Within ML, algorithms are used to generalize situations beyond those seen in training data and then complete tasks without further guidance from people. A good example is U-M GPT. The large language models (LLMs) accessible via U-M GPT are trained with millions of diverse examples. “The goal is to train the models to reliably predict, translate, or generate something.” 

“Data is any information that can be formatted and fed into an algorithm that can be used for some task, including journal articles, chats, numbers, videos, audio, and texts,” said Brown, Algorithms can be trained to perform tasks using these real-world data. 

Natural Language Processing is a branch of artificial intelligence that helps computers understand and generate human language in a way that is both meaningful and useful to humans. NLP teaches computers to understand languages and then respond so that humans can understand, and even accounting for when rich context language is used. 

“NLP is highly interdisciplinary, and involves multiple fields, such as computer science, linguistics, philosophy, cognitive science, statistics, mathematics, etc.,” said Chai.

Examples of NLP are everywhere: when you ask Siri for directions, or when Google efficiently completes your half-typed query, or even when you get suggested replies in your email. 

Ultimately NLP, along with AI, can be used to make interactions between humans and machines as natural and as easy as possible. 

A lot of data is needed to train the models

Dr. Chai and her team use large language models, a lot of data, and computing resources. These models take longer to train and are harder to interpret. Brown says, “The state of the art, groundbreaking work tends to be in this area.” 

Dr. Chai uses deep learning algorithms that make predictions about what the next part of the task or conversation is. “For example, they use deep learning and the transformer architecture to enable embodied agents to learn how new words are connected to the physical environment, to follow human language instructions, and to collaborate with humans to come up with a shared plan,” Brown explains.

The technology that supports this work

To accomplish her work, Dr. Chai uses the Great Lakes High-Performance Computing Cluster and Turbo Research Storage, both of which are managed by U-M’s Advanced Research Computing Group (ARC) in Information and Technology Services. She has 16 GPUs on Great Lakes at the ready, with the option to use more at any given time. 

A GPU, or Graphics Processing Unit, is a piece of computer equipment that is good at displaying pictures, animations, and videos on your screen. The GPU is especially adept at quickly creating and manipulating images. Traditionally, GPUs were used for video games and professional design software where detailed graphics were necessary. But more recently, researchers including Dr. Chai discovered that GPUs are also good at handling many simple tasks at the same time. This includes tasks like scientific simulations and AI training where a lot of calculations need to be done in parallel (which is perfect for training large language models).

“GPUs are popular for deep learning, and we will continue to get more and better GPUs in the future. There is a demand, and we will continue supporting this technology so that deep learning can continue to grow,” said Brock Palen, ITS Advanced Research Computing director. 

Chai and her team also leveraged 29 terabytes of the Turbo Research Storage service at ARC. NLP benefits from the high-capacity, reliable, secure, and fast storage solution. Turbo enables investigators across the university to store and access data needed for their research via Great Lakes. 

Great Lakes HPC in the classroom 

ARC offers classroom use of high-performance computing cluster resources on the Great Lakes High-Performance Computing Cluster

Dr. Chai regularly leverages this resource. “Over 300 students have benefited from this experience. We have homework that requires the use of the Great Lakes, e.g., having students learn how to conduct experiments in a managed job-scheduling system like SLURM. This will benefit them in the future if they engage in any compute-intensive R&D (research and development).

“For my NLP class, I request Great Lakes access for my students so they have the ability to develop some meaningful final projects. We also use the Great Lakes HPC resources to study the reproducibility for NLP beginners,” said Chai. A gallery is available for many of the student projects.

The UMRCP defrays costs

The U-M Research Computing Package is a set of cost-sharing allocations offered by ITS ARC, and they are available upon request. Other units offer additional cost-sharing to researchers. Chai said, “We typically use the nodes owned by my group for research projects that require intensive, large-scale GPU model training. We use the UMRCP for less intensive tasks, thereby extending the budgetary impact of the allocations.” 

New GPU Offering

By | Uncategorized

We are pleased to introduce NVIDIA’s Multi-Instance GPU (MIG) technology to Great Lakes, featuring a total of 16 GPUs. This development promises to optimize your computing experience.

How MIG Works:

NVIDIA’s Multi-Instance GPU (MIG) technology divides our 8 GPUs into 16 multiple isolated instances, each behaving as an independent GPU with dedicated compute resources. This partitioning allows for efficient allocation of GPU resources, enhancing your computing experience. 

Key Benefits and Limitations:

  • Efficient Resource Allocation: MIG’s partitioning ensures that your tasks receive dedicated GPU resources, avoiding resource contention and enhancing efficiency.
  • Enhanced Scalability: Run multiple GPU workloads concurrently without conflicts, simplifying project scaling.
  • Flexibility: Customize GPU instances to match your application requirements, optimizing performance and cost-effectiveness.
  • MIG is only intended for single-slice jobs, and a single process cannot run across multiple devices. Slurm will only allow jobs requesting a single GPU.

Getting Started:

To access the nodes equipped with MIG technology, use the Slurm partition called “gpu_mig40” when submitting your job requests. The partition is named with the amount of memory each GPU has available.

Important Notice: Please be aware that each job can only utilize a single GPU. If your jobs needs more that 1 GPU, use the “gpu” or “spgpu” partition, depending on your GPU needs.

Example Slurm Job Submission:

sbatch –partition=gpu_mig40 –gres=gpu:1 your_job_script.sh

If your job could run on either the “gpu_mig40” or “gpu” partition, you can specify both, and the scheduler will schedule your job on either partition.

sbatch –partition=gpu_mig40,gpu –gres=gpu:1 your_job_script.sh

For questions or support requests, please contact our team at arc-support@umich.edu