Pinned Repositories
ArtBecomeHuman
Training a convolutional neural net to distinguish AI-generated art from human made art.
ASKiT
Stanford CS224N Final Project. A text-based multi-hop reasoning question-answer NLP model to address the research question: "What if language models could ask their own questions?"
AttenPi
An end-to-end differentiable hierarchical reinforcement learning agent based on continuous sub-policy attention.
DimensionalCUDA
A set of matrix and vector data structures for easy multi-dimensional data manipulation with CUDA.
LiNet
A transfer-function-based recurrent neural net that can backpropagate forward through time.
MonArc
A practical method for training energy-based language models.
NoiseSearch
Inference-time search algorithms to optimize the 'noise' in diffusion-based text-to-image models for arbitrary objective optimization.
Orbit-Propogator
Earth centric orbit propagation tool. Built from scratch in python.
P-you-
Two-factor password authentication using biometric keystroke data.
Shors-Algorithm
An implementation of Shor's quantum factoring algorithm on the number 15.
aklein4's Repositories
aklein4/ASKiT
Stanford CS224N Final Project. A text-based multi-hop reasoning question-answer NLP model to address the research question: "What if language models could ask their own questions?"
aklein4/ArtBecomeHuman
Training a convolutional neural net to distinguish AI-generated art from human made art.
aklein4/AttenPi
An end-to-end differentiable hierarchical reinforcement learning agent based on continuous sub-policy attention.
aklein4/Orbit-Propogator
Earth centric orbit propagation tool. Built from scratch in python.
aklein4/P-you-
Two-factor password authentication using biometric keystroke data.
aklein4/ArithSearch
A research repo for computation graph optimization using metaheuristic methods.
aklein4/ArtBecomeHumanDeliverables
Presentation Poster from CS229 final project.
aklein4/BoltFlow
Efficient distillation of diffusion-based text-to-image models using only a teacher model and prompts (no training images).
aklein4/DimensionalCUDA
A set of matrix and vector data structures for easy multi-dimensional data manipulation with CUDA.
aklein4/img2vec
An experiment to see if an image classification algorithm can independently learn features found in word vectors.
aklein4/LiNet
A transfer-function-based recurrent neural net that can backpropagate forward through time.
aklein4/MonArc
A practical method for training energy-based language models.
aklein4/NoiseSearch
Inference-time search algorithms to optimize the 'noise' in diffusion-based text-to-image models for arbitrary objective optimization.
aklein4/Shors-Algorithm
An implementation of Shor's quantum factoring algorithm on the number 15.
aklein4/SkillBasis
Stanford CS224R project: Skills as Gradients in a Latent Field
aklein4/Supermileage-Car
Plans for Mauston Super Mileage Car #3, 2019-2020. Includes models, drawings, and samples images.
aklein4/ArcImg
Residual energy models for image classification.
aklein4/ArcLM
Building an LLM with higher logit precision using a single-token energy model.
aklein4/git-filter-repo
Quickly rewrite git repository history (filter-branch replacement)
aklein4/ImageSearch
Using inference-time search methods to improve and control diffusion-based image generation.
aklein4/paladin
A deep learning experiment to combine the principles of LDMs and autoregressive LLMs to create a language model that can think ahead.