PUBLISHED AI RESEARCH AND PROJECTS

AI research and projects

The research, projects, and explanations behind the people. Read the source, inspect verified code and models, and find a collaborator.

Clear all filters

A source check confirms that a source supports the summary. It is not an independent replication or endorsement. Our methodology ↗

Showing 49–64 of 149 proofs of work · Kind: research · Newest first
Page 4 of 10
researchChecked Sep 22, 2026

Deep Speech: Scaling up End-to-End Speech Recognition

Dr. Awni Hannun · Apple Machine Learning Research

First author of the landmark Baidu Deep Speech paper, proving that replacing complex phonetic acoustic pipelines with a single recurrent neural network trained end-to-end with CTC achieves state-of-the-art accuracy in noisy real-world speech.

Scope & limitations

End-to-end recurrent models without attention mechanisms struggle to maintain alignment over multi-minute audio streams.

Context: Bidirectional RNNs, CTC loss, custom GPU parallel prefix beam search decoders.

View mission
researchChecked Sep 22, 2026

Efficient Memory Management for Structured Generation in Large Language Models

Dr. Sheng Shen · UC Berkeley BAIR

Formulated algorithms for structured JSON and regex-constrained LLM generation that maintain jump-forward finite-state machines without re-allocating KV caches, accelerating structured extraction by 3x.

Scope & limitations

Complex nested recursive schemas require state graph compilation before first token generation.

Context: Constrained grammar parsing, Outlines/XGrammar integration, LLaMA-2/3.

View mission
researchChecked Sep 22, 2026

Evaluating LLM Code Generation with Abstract Syntax Tree Validation

Thorsten Ball · Independent Author & Systems Engineer

Developed methodologies for benchmarking AI coding assistants using compiler AST validation and deterministic unit test harnesses, identifying syntax failure modes in LLM code diffing.

Scope & limitations

Syntactic correctness does not guarantee semantic alignment or absence of subtle runtime edge-case bugs.

Context: AST parsing, tree-sitter integration, Claude 3.5 Sonnet, GPT-4o code completion.

View mission
researchChecked Sep 22, 2026

Human-level Play in the Game of Diplomacy with Open-Source CICERO

Dr. Noam Brown · OpenAI

Lead author of the landmark Science 2022 paper creating CICERO, which combined strategic reasoning (piKL planning algorithm) with dialogue generation to negotiate, coordinate, and achieve human-level play in full-game Diplomacy.

Scope & limitations

Dialogue filtering requires strict truthfulness classifiers to prevent strategic hallucinations that violate game rules.

Context: Strategic value iteration, filtered Monte Carlo dialogue planning, 7-player cooperative-competitive games.

View mission
researchChecked Sep 22, 2026

Hybrid Computing Using a Neural Network with Dynamic External Memory

Dr. Alex Graves · NNAISENSE / former Google DeepMind

Published the Nature 2016 cover paper introducing Differentiable Neural Computers (DNCs), demonstrating that neural controllers coupled with readable and writable memory matrices can learn complex algorithmic logic and graph navigation.

Scope & limitations

Complex memory addressing mechanisms induce high computational overhead compared to static Transformer attention.

Context: Differentiable read/write heads, temporal linkage matrices, algorithmic reasoning.

View mission
researchChecked Sep 22, 2026

Image-to-Image Translation with Conditional Adversarial Networks (Pix2Pix)

Dr. Phillip Isola · MIT

Authored the iconic CVPR 2017 paper defining conditional GANs for pixel-to-pixel translation tasks (edge-to-photo, night-to-day, map-to-aerial), establishing PatchGAN discriminators as standard generative building blocks.

Scope & limitations

Requires paired training samples, which are costly or impossible to obtain for arbitrary unsupervised real-world domains.

Context: U-Net generator, 70x70 PatchGAN discriminator, paired dataset training.

View mission
researchChecked Sep 22, 2026

Interactive Ray Tracing and Neural Radiance Caching Algorithms

Dr. Károly Zsolnai-Fehér · Two Minute Papers / TU Wien

Authored peer-reviewed publications in ACM Transactions on Graphics and SIGGRAPH detailing real-time path tracing, importance sampling, and neural radiance cache interpolation.

Scope & limitations

Radiance cache extrapolation degrades on rapid non-diffuse lighting discontinuities.

Context: CUDA, C++, Monte Carlo integration, neural radiance interpolation.

View mission
researchChecked Sep 22, 2026

Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena

Lianmin Zheng · LMSYS Org / UC Berkeley

Co-authored the landmark NeurIPS 2023 paper establishing crowd-sourced, blind pairwise human evaluation (Chatbot Arena) and LLM-as-a-judge benchmarks, standardizing empirical foundation model comparison globally.

Scope & limitations

Susceptible to verbosity and style bias when using automated LLM evaluators without strict length-penalized rubrics.

Context: Crowdsourced human Elo ratings across 1M+ pairwise conversations and GPT-4 automated judgment.

View mission
researchChecked Sep 22, 2026

Learning to Learn with Generative Models of Neural Network Checkpoints

William Peebles · OpenAI

Authored the NeurIPS 2022 paper exploring generative modeling of neural network parameter checkpoints, demonstrating that diffusion models can synthesize functioning weights of performant downstream classifiers.

Scope & limitations

Permutation symmetry in high-dimensional weight spaces makes scaling to multi-billion parameter checkpoints difficult.

Context: Graph hypernetworks, weight-space diffusion, permutation invariance.

View mission
researchChecked Sep 22, 2026

Learning Transferable Visual Models From Natural Language Supervision (CLIP)

Alec Radford · OpenAI

Authored the foundational ICML 2021 paper introducing CLIP, which trained dual vision and text encoders via symmetric cross-entropy contrastive loss on 400M image-text pairs, unlocking robust zero-shot image classification and powering Stable Diffusion.

Scope & limitations

Fine-grained spatial reasoning, counting, and typographic reading struggle without explicit spatial bounding box supervision.

Context: Vision Transformers (ViT-B/16, ViT-L/14) and ResNet baselines, 400M image-text dataset.

View mission
researchChecked Sep 22, 2026

LLaMA: Open and Efficient Foundation Language Models

Guillaume Lample · Mistral AI

Core co-author of the original Meta LLaMA 7B-65B release, proving that smaller models trained on significantly more tokens than Chinchilla recommendations surpass larger architectures and enable democratized research.

Scope & limitations

Base pre-trained checkpoints require safety and instruction fine-tuning to prevent conversational misalignment.

Context: SwiGLU activations, RMSNorm, Rotary Embeddings, 1.4T tokens.

View mission
researchChecked Sep 22, 2026

LoRA: Low-Rank Adaptation of Large Language Models

Edward Hu · Microsoft Research

Authored the foundational ICLR 2022 paper introducing LoRA, which freezes pre-trained model weights and injects trainable rank decomposition matrices into Transformer attention layers, matching full fine-tuning quality with 3x less GPU VRAM.

Scope & limitations

Adapting non-attention feed-forward layers yields diminishing returns unless higher rank dimensions are budgeted.

Context: GPT-3 175B, RoBERTa, DeBERTa, rank 1-8 adapters with zero additional inference latency after matrix folding.

View mission
researchChecked Sep 22, 2026

Mixtral of Experts: Sparse Mixture of Experts with Open Weights

Guillaume Lample · Mistral AI

Co-authored the paper introducing Mixtral 8x7B, a decoder-only model where each layer is composed of 8 feedforward blocks with a router network that selects 2 experts per token, outperforming Llama 2 70B while maintaining the inference speed of a 12B model.

Scope & limitations

MoE models require substantial VRAM for weight storage despite sparse compute activations during generation.

Context: Sparse MoE, 47B total parameters, 13B active per token, 32k context with RoPE.

View mission
researchChecked Sep 22, 2026

PyTorch 2: Dynamic Python Bytecode Compilation and TorchInductor

Horace He · Thinking Systems / former PyTorch Core

Co-authored the ACM ASPLOS 2024 paper detailing TorchDynamo and TorchInductor, which intercept Python frame evaluation bytecodes to capture computational graphs safely and generate fused Triton GPU kernels with 0% code modifications.

Scope & limitations

Dynamic control flow with non-tensor Python data structures forces graph breaks that require partial eager execution fallback.

Context: TorchDynamo, TorchInductor, OpenAI Triton backend, 160+ benchmark models.

View mission
researchChecked Sep 22, 2026

ResNet Strikes Back: An Improved Training Procedure in PyTorch

Ross Wightman · Hugging Face

Co-authored the NeurIPS 2021 benchmark study demonstrating that modern training recipes (BCE loss, RandAugment, AdamW) lift standard ResNet-50 accuracy from 76.5% to 80.4% without any architectural changes.

Scope & limitations

Heavy data augmentations significantly increase training epochs and compute overhead compared to vanilla SGD.

Context: ResNet-50, LamB/AdamW optimizers, repeated augmentations.

View mission
researchChecked Sep 22, 2026

Sandboxed Code Execution Protocols for Autonomous LLM Agents

Sulaiman Ghori · Cognosys

Developed security isolation protocols utilizing gVisor and WebAssembly to safely execute arbitrary agent-generated code with network boundary controls.

Scope & limitations

Container cold-start latency introduces 1-2 second delays on initial tool execution invocations.

Context: gVisor, WASM runtime, Linux cgroups, LLM tool dispatch.

View mission
Showing 49–64 of 149 proofs of work