Papers by Kurt Keutzer
LLM2LLM: Boosting LLMs with Novel Iterative Data Enhancement (2024.findings-acl)
Copied to clipboard
Nicholas Lee, Thanakul Wattanawong, Sehoon Kim, Karttikeya Mangalam, Sheng Shen, Gopala Anumanchipalli, Michael Mahoney, Kurt Keutzer, Amir Gholami
| Challenge: | Pretrained large language models are currently state-of-the-art for solving most tasks . however, many of them are in the low-data regime, making fine-tuning challenging . a new data augmentation strategy uses a teacher LLM to augment a small seed dataset . |
| Approach: | They propose a targeted and iterative data augmentation strategy that augments a teacher LLM to fine-tune a small seed dataset by adding additional data. |
| Outcome: | The proposed approach outperforms fine-tuning and other data augmentation strategies on a small seed dataset. |
S*: Test Time Scaling for Code Generation (2025.findings-emnlp)
Copied to clipboard
Dacheng Li, Shiyi Cao, Chengkun Cao, Xiuyu Li, Shangyin Tan, Kurt Keutzer, Jiarong Xing, Joseph E. Gonzalez, Ion Stoica
| Challenge: | S* is the first hybrid test-time scaling framework that significantly improves the coverage and selection accuracy of generated code. |
| Approach: | They propose a hybrid test-time scaling framework that augments parallel scaling with sequential scaling to further increase the performance. |
| Outcome: | The proposed framework outperforms existing scaling approaches in large-scale modeling and reasoning models. |
LLoCO: Learning Long Contexts Offline (2024.emnlp-main)
Copied to clipboard
Sijun Tan, Xiuyu Li, Shishir G Patil, Ziyang Wu, Tianjun Zhang, Kurt Keutzer, Joseph Gonzalez, Raluca Popa
| Challenge: | Large language models are still unable to handle long contexts due to the quadratic computational and memory overhead of the self-attention mechanism and the substantial KV cache sizes during generation. |
| Approach: | They propose a method to learn contexts offline through context compression and in-domain parameter-efficient finetuning with LoRA. |
| Outcome: | The proposed model outperforms in-context learning while using 30 fewer tokens during inference and significantly reduces the cost of long document question answering. |
What’s Hidden in a One-layer Randomly Weighted Transformer? (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing subnetworks of one-layer randomly weighted neural networks can achieve impressive performance without changing initializations. |
| Approach: | They find subnetworks within one-layer randomly weighted neural networks that can achieve impressive performance without ever modifying the initializations. |
| Outcome: | The proposed subnetworks match 98%/92% of the performance of a trained Transformersmall/base on IWSLT14/WMT14. |
Simple and Effective Input Reformulations for Translation (2023.emnlp-main)
Copied to clipboard
| Challenge: | Foundation language models learn from their finetuning input context in different ways. |
| Approach: | They propose three different data efficient techniques to improve translation performance . they reformulate inputs during finetuning for challenging translation tasks . |
| Outcome: | The proposed techniques show significant improvements on the Flores200 translation benchmark. |
Squeezed Attention: Accelerating Long Context Length LLM Inference (2025.acl-long)
Copied to clipboard
Coleman Richard Charles Hooper, Sehoon Kim, Hiva Mohammadzadeh, Monishwaran Maheswaran, Sebastian Zhao, June Paik, Michael W. Mahoney, Kurt Keutzer, Amir Gholami
| Challenge: | Emerging Large Language Models require long input context to perform complex tasks. |
| Approach: | They propose an algorithm to reduce the complexity of attention with respect to the fixed context length. |
| Outcome: | The proposed method reduces the complexity of attention from linear to logarithmic with respect to the fixed context length. |
Scaling Vision-Language Models with Sparse Mixture of Experts (2023.findings-emnlp)
Copied to clipboard
| Challenge: | a study explores the effectiveness of mixture-of-experts (MoE) techniques in scaling vision-language models . alayrac and colleagues demonstrate the effectiveness and performance of MoE in scaling VLMs . |
| Approach: | They propose to use sparsely-gated mixture-of-experts techniques to scale vision-language models . they show that MoE can achieve state-of the-art performance over dense models a range of benchmarks . |
| Outcome: | The proposed approach achieves state-of-the-art performance over dense models of equivalent computational cost. |
MAgIC: Investigation of Large Language Model Powered Multi-Agent in Cognition, Adaptability, Rationality and Collaboration (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) have advanced natural language processing, demonstrating exceptional reasoning, tool usage, and memory capabilities. |
| Approach: | They propose a competition-based benchmark framework specifically designed to assess LLMs within multi-agent environments. |
| Outcome: | The proposed framework enhances the LLMs’ abilities in navigating complex social and cognitive dimensions by over threefold between the strongest and weakest LLM models. |
Reservoir Transformers (2021.acl-long)
Copied to clipboard
| Challenge: | Using random initialization, we show that some transformers obtain impressive performance even when some of the layers are frozen. |
| Approach: | They propose to freeze transformer layers and use them to improve performance . they find that the transformers obtain impressive performance even when some of the layers are randomly initialized and never updated. |
| Outcome: | The proposed model improves on translation and language modelling tasks even when some layers are frozen. |
Aligning Large Multimodal Models with Factually Augmented RLHF (2024.findings-acl)
Copied to clipboard
Zhiqing Sun, Sheng Shen, Shengcao Cao, Haotian Liu, Chunyuan Li, Yikang Shen, Chuang Gan, Liangyan Gui, Yu-Xiong Wang, Yiming Yang, Kurt Keutzer, Trevor Darrell
| Challenge: | Large Multimodal Models (LMMs) are built across modalities and the misalignment between two modality can result in "hallucination" . developing LMMs faces challenges such as a lack of data and a limited number of data sets. |
| Approach: | They propose a new algorithm that augments the reward model with additional factual information such as image captions and ground-truth multi-choice options. |
| Outcome: | The proposed approach improves on the LLaVA-Bench dataset with the 96% performance level of the text-only GPT-4 and an improvement of 60% on MMHAL-BENCH over other baselines. |
TinyAgent: Function Calling at the Edge (2024.emnlp-demo)
Copied to clipboard
Lutfi Erdogan, Nicholas Lee, Siddharth Jha, Sehoon Kim, Ryan Tabrizi, Suhong Moon, Coleman Hooper, Gopala Anumanchipalli, Kurt Keutzer, Amir Gholami
| Challenge: | Recent large language models (LLMs) have enabled the development of advanced agentic systems that can integrate various tools and APIs to fulfill user queries. |
| Approach: | They propose an end-to-end framework for training and deploying task-specific small language model agents capable of function calling for driving agentic systems at the edge. |
| Outcome: | The proposed model outperforms existing models by reducing the input prompt length and quantizing the inference speed. |
One Model is All You Need: ByT5-Sanskrit, a Unified Model for Sanskrit NLP Tasks (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Morphologically rich languages are notoriously challenging to process for downstream NLP applications. |
| Approach: | They propose a pretrained model for NLP applications involving the morphologically rich language Sanskrit that outperforms previous models by a considerable margin. |
| Outcome: | The proposed model outperforms tokenized models on established Sanskrit word segmentation tasks and matches the current best lexicon-based model. |