-
Liam Thomas Conroy
22 Year old Mechanical Engineering Graduate student from Binghamton University. I like fast cars, but not on the road.
https://nanohub.org/members/461182
-
W43 - Efficient and Reliable Deep Learning at Scale: Hardware and Software
Online Presentations | 04 Sep 2024 | Contributor(s):: Yiran Chen
This talk discusses the importance of hardware/software co-design in the development of AI computing systems. We first use resistive memory based Neural Network (NN) accelerators to illustrate the design philosophy of heterogeneous AI computing systems, and then present several...
-
W44 - Securing Neural Networks Against Side-Channel Attacks with Hardware Masking
Online Presentations | 03 Sep 2024 | Contributor(s):: Aydin Aysu
In this talk, I will explain different mechanisms to mask neural networks in hardware and describe related opportunities and challenges. I will first discuss how a straightforward masking adaptation leaks side-channel information on neural networks and how to address this...
-
W47 - Fault Criticality Assessment in AI Accelerators
Online Presentations | 03 Sep 2024 | Contributor(s):: Krishnendu Chakrabarty
This presentation presents a neural-twin framework for analyzing fault criticality with a negligible amount of ground-truth data. A recently proposed misclassification-driven training algorithm will be used to sensitize and identify biases that are critical to the functioning of the accelerator...
-
W46 - DisruptNet: Integrity Breach of Deep Neural Network Execution
Online Presentations | 03 Sep 2024 | Contributor(s):: Yunsi Fei
This talk focuses on several new active attacks of deep learning accelerators on different platforms, FPGA and GPU, resulting in image misclassification and integrity breach of deep neural network execution. Our new vector of attacks are first of their kind and reveal a largely under-explored...
-
Materials Informatics Lecture 30: Bayes Theorem and Naive Bayes Classifier
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Welcome back to our Materials Informatics series! In this video, Taylor Sparks from the University of Utah delves into the fundamentals of Bayes Theorem and Naive Bayes classifiers, essential concepts in probabilistic machine learning. In this video, we cover: An introduction to Bayesian...
-
Materials Informatics Lecture 30a: Coding a Naive Bayes Classifier by Hand
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Welcome back to our Materials Informatics series! In this video, we continue exploring Bayesian and probabilistic machine learning by diving deeper into Naive Bayes classification.In this video, we cover:A brief recap of Naive Bayes, highlighting its simplicity and the assumption of feature...
-
Materials Informatics Lecture 31: Gaussian Processes
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Welcome back to our Materials Informatics playlist! In this video, we dive into the fascinating world of Gaussian Processes, building on our previous discussion about Naive Bayes theorem. Gaussian Processes are a powerful tool in probabilistic machine learning, especially for modeling and...
-
Materials Informatics Lecture 32: Bayesian Optimization
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Welcome back to our Materials Informatics series! In today's episode, we delve into Bayesian Optimization, a critical tool for incrementally improving processes and designs in materials research. Bayesian Optimization leverages Bayes' theorem to make informed decisions with minimal data,...
-
Materials Informatics Lecture 33: Large Language Models in Materials Science
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Welcome back to our Materials Informatics playlist! In today's episode, we explore the transformative impact of Large Language Models (LLMs) on materials science. David Sparks from the University of Utah's Materials Science and Engineering department takes us through the exciting...
-
Materials Informatics Lecture 27: Variational Autoencoders
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Generative machine learning models have the potential to allow us to move beyond screening to true materials discovery. Generative adversarial networks (GANs) are one powerful tool and variational autoencoders (VAEs) are another. This video descrbies autoencoders, latent space, reparameterization...
-
Materials Informatics Lecture 28: Image Segmentation
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Welcome back to our Materials Informatics playlist! In our previous video, we explored generative models like GANs. Today, we’re taking a brief interlude to discuss the fascinating world of image segmentation because tools from image segmentation are crucial to our next generative model!In...
-
Materials Informatics Lecture 29: Diffusion Models
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Welcome back to our Materials Informatics series! In this video, we're diving into the fascinating world of diffusion models, the cutting-edge technology in generative models that is outperforming GANs and VAEs.In this video, we cover:Introduction to diffusion models and their superiority in...
-
Materials Informatics Lecture 26: Generative Adversarial Networks
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
In this video we cover the first video on generative machine learning models: the generative adversarial network. We describe what they are, how they work, why they are so hard to train and more.
-
Materials Informatics Lecture 26a: Application of GANs in Materials Science
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
With our introduction to GANs complete, we now do a survey of some of the places where we see them applied in materials science research. From inverse design of new materials and molecules to data augmentation we see the power of these generative ML models.
-
Materials Informatics Lecture 22: Neural Networks
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Neural networks have been the come-back kids of machine learning. They were invented in the 80s, but surged back into the limelight in the late 90s to become a powerful tool at the heart of deep learning. This video describes what neural networks are, how they work in materials informatics,
-
Materials Informatics Lecture 23: Convolutional Neural Networks
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
Vanilla neural networks are powerful, but convolutional neural networks are truly revolutionary! Instead of constructing features by hand, a convolutional neural network can extract features on its own! It does this through convolutional layers and then reduces dimensions for faster computing...
-
Materials Informatics Lecture 24: Recurrent Neural Networks
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
How do we deal with sequential data? How do we make a machine learning model pay attention to data where order matters? A big innovation came with the development of recurrent neural networks and their modern versions (LSTM and GRU).
-
Materials Informatics Lecture 25: Transformers
Online Presentations | 13 Aug 2024 | Contributor(s):: Taylor Sparks
As good as recurrent networks are, they still face fundamental limitations. Vanishing gradient makes it so RNNs can only look back ~100 data points in a series or ~100 in an LSTM or GRU. Transformers let us learn how all inputs are related to all other inputs through the novel "multi-head...
-
nanoHUB-U: Mathematics of Waves - Visualized with Neural Networks
Courses | 26 Jun 2024 | Contributor(s):: Supriyo Datta
All around us, information is carried by waves of many kinds, like acoustic, electromagnetic, gravitational or probability waves. Presented is a unifying view of all wave phenomena using neural networks to visualize the underlying mathematics.