Skip to main content

Cyberinfrastructure

aiicer-logo

Pioneering real-world AI solutions, building talent, and driving regional transformation with global impact.

Visualization Wall

A high-resolution, large-format display environment for scientific data exploration, multidisciplinary collaboration, and high-profile presentations.

High-Performance Computing (HPC)

Provides high-performance computing capacity for running AI models, scientific simulations, and large-scale data processing workloads.

Under Construction

Training and support in the use of the following tools and platforms:

Tool/Software Description/Use Case Link
PyTorch Deep learning framework for building and training neural networks https://pytorch.org/
TensorFlow End-to-end open-source platform for machine learning https://www.tensorflow.org/
FLOSIC Software for self-interaction correction in density functional theory https://www.flosic.org/
Docker Containerization platform for deploying and managing applications https://www.docker.com/
Podman Daemonless container engine for developing, managing, and running containers https://podman.io/
SLURM Workload manager for high-performance computing environments https://slurm.schedmd.com/documentation.html
Anaconda Free data science platform for Python and R, offering a simplified way to manage packages, environments, and projects, especially for AI, machine learning, and scientific computing https://anaconda.org/
NVIDIA NGC AI-driven platform and marketplace providing GPU-accelerated software, containers, models, and SDKs, optimized for deep learning and high-performance computing (HPC) https://catalog.ngc.nvidia.com/
A table describing the tools/software that get used for support and training.

Multimodal open-source models

Models that integrate multiple sources of information (e.g., images, text, audio) to more effectively address machine-learning tasks, including the fusion of data from multiple sensors and cross-modal analysis.

OpenCLIP


Category: Vision-Language encoder

BLIP-2
(LAVIS)


Category: Vision → Text / Instruction tuning

LLaVA


Category: Multimodal LLM (image+text)

Stable Diffusion


Category: Text → Image

Whisper


Category: Speech → Text

OpenMMLab


Category: Vision-Language encoder

LLama


Category: Language understanding, reasoning, and text generation

Scientific Workflow

Connect data, models, and simulations into reproducible AI-enabled research pipelines.

SWIM (Scientific Workflow Integration & Management)

A platform that connects scientific models with the data they need, providing cloud hosting and an easy-to-use interface for integration, execution and interpretation of water models.

Active Models in SWIM

Middle Rio Grande Water Balance Model

Simulates regional water flows, reservoir operations, and groundwater use across the Middle Rio Grande under historical and projected climate conditions.

View Repository Integration type: Data-to-Model

Middle Rio Grande Hydroeconomic Optimization Model

Optimizes reservoir operations and water allocation decisions under economic and institutional constraints.

View Repository Integration type: Data-to-Model

Automating Multivariable Workflow Composition for Model-to-Model Integration

Model-to-Model integration automatically connects two scientific models by using the outputs of one as inputs to another. This enables seamless multi-model workflows for more complete and realistic scenario analysis.

Read Paper Integration type: Model-to-Model

For questions or to learn more, please contact aaii@utep.edu.