Incoherence Processing

Incoherence processing addresses the challenge of improving the consistency and logical flow within data, whether it's text, images, or the internal representations of large language models (LLMs). Current research focuses on developing algorithms and model architectures, such as those employing vector quantization, trellis coded quantization, and Hadamard transforms, to detect and correct incoherence, particularly in the context of LLM quantization and improving the quality of machine-generated text. These advancements aim to enhance the efficiency and performance of LLMs while also improving the quality of machine-generated content, impacting fields ranging from natural language processing to image generation. The development of robust incoherence processing techniques is crucial for advancing the capabilities and reliability of artificial intelligence systems.

Papers

June 28, 2024

DECOR: Improving Coherence in L2 English Writing with a Novel Benchmark for Incoherence Detection, Reasoning, and Rewriting
Xuanming Zhang, Anthony Diaz, Zixun Chen, Qingyang Wu, Kun Qian, Erik Voss, Zhou Yu
Complex Reasoning Evaluation Method Discourse Coherence Novel Benchmark Coherence Evaluation L2 Speech Incoherence Processing

June 17, 2024

QTIP: Quantization with Trellises and Incoherence Processing
Albert Tseng, Qingyao Sun, David Hou, Christopher De Sa
Quantization Operator Post Training Quantization Vector Quantization Novel Encoding Extreme Quantization Tree Structure Quantization Performance Incoherence Processing

February 6, 2024

QuIP#: Even Better LLM Quantization with Hadamard Incoherence and Lattice Codebooks
Albert Tseng, Jerry Chee, Qingyao Sun, Volodymyr Kuleshov, Christopher De Sa
Post Training Quantization Vector Quantization Codebook Learning LLM Quantization Hadamard Layer Incoherence Processing

July 25, 2023

QuIP: 2-Bit Quantization of Large Language Models With Guarantees
Jerry Chee, Yaohui Cai, Volodymyr Kuleshov, Christopher De Sa
Large Language Model Quantization Operator Formal Guarantee LLM Quantization Parameter Quantization Incoherence Processing

April 6, 2023

Towards Coherent Image Inpainting Using Denoising Diffusion Implicit Models
Guanhua Zhang, Jiabao Ji, Yang Zhang, Mo Yu, Tommi Jaakkola, Shiyu Chang
Diffusion Model Diffusion Based Method Posterior Distribution Denoising Diffusion Implicit Model Clean Image Incoherence Processing

Incoherence Processing

Papers

DECOR: Improving Coherence in L2 English Writing with a Novel Benchmark for Incoherence Detection, Reasoning, and Rewriting

QTIP: Quantization with Trellises and Incoherence Processing

QuIP#: Even Better LLM Quantization with Hadamard Incoherence and Lattice Codebooks

QuIP: 2-Bit Quantization of Large Language Models With Guarantees

Towards Coherent Image Inpainting Using Denoising Diffusion Implicit Models