Cheng Li

Zhewei Yao, Xiaoxia Wu, Cheng Li et al.

Post-training quantization (PTQ) has emerged as a promising technique for mitigating memory consumption and computational costs in large language models (LLMs). However, a systematic examination of various quantization schemes, model families, and quantization bit precision has been absent from the literature. In this paper, we conduct a comprehensive analysis of these factors by investigating the effects of PTQ on weight-only, activation-only, and weight-and-activation quantization using diverse methods such as round-to-nearest (RTN), GPTQ, ZeroQuant, and their variants. We apply these methods to two distinct model families with parameters ranging from 125M to 176B. Our contributions include: (1) a sensitivity analysis revealing that activation quantization is generally more susceptible to weight quantization, with smaller models often outperforming larger models in terms of activation quantization; (2) an evaluation and comparison of existing PTQ methods to optimize model size reduction while minimizing the impact on accuracy, revealing that none of the current methods can achieve the original model quality for quantization with either INT4-weight or INT4-weight-and-INT8-activation; (3) based on these insights, we propose an optimized method called Low-Rank Compensation (LoRC), which employs low-rank matrices to enhance model quality recovery with a minimal increase in model size.

Zhewei Yao, Xiaoxia Wu, Conglong Li et al.

Large-scale transformer models have become the de-facto architectures for various machine learning applications, e.g., CV and NLP. However, those large models also introduce prohibitive training costs. To mitigate this issue, we propose a novel random and layerwise token dropping method (random-LTD), which skips the computation of a subset of the input tokens at all middle layers. Particularly, random-LTD achieves considerable speedups and comparable accuracy as the standard training baseline. Compared to other token dropping methods, random-LTD does not require (1) any importance score-based metrics, (2) any special token treatment (e.g., [CLS]), and (3) many layers in full sequence length training except the first and the last layers. Besides, a new LayerToken learning rate schedule is proposed for pretraining problems that resolve the heavy tuning requirement for our proposed training mechanism. Finally, we demonstrate that random-LTD can be applied to broader applications, including GPT and BERT pretraining as well as ViT and GPT finetuning tasks. Our results show that random-LTD can save about 33.3% theoretical compute cost and 25.6% wall-clock training time while achieving similar zero-shot evaluations on GPT-31.3B as compared to baseline.

Wei Xie, Bo Wang, Cheng Li et al.

While biomanufacturing plays a significant role in supporting the economy and ensuring public health, it faces critical challenges, including complexity, high variability, lengthy lead time, and very limited process data, especially for personalized new cell and gene biotherapeutics. Driven by these challenges, we propose an interpretable semantic bioprocess probabilistic knowledge graph and develop a game theory based risk and sensitivity analyses for production process to facilitate quality-by-design and stability control. Specifically, by exploring the causal relationships and interactions of critical process parameters and quality attributes (CPPs/CQAs), we create a Bayesian network based probabilistic knowledge graph characterizing the complex causal interdependencies of all factors. Then, we introduce a Shapley value based sensitivity analysis, which can correctly quantify the variation contribution from each input factor on the outputs (i.e., productivity, product quality). Since the bioprocess model coefficients are learned from limited process observations, we derive the Bayesian posterior distribution to quantify model uncertainty and further develop the Shapley value based sensitivity analysis to evaluate the impact of estimation uncertainty from each set of model coefficients. Therefore, the proposed bioprocess risk and sensitivity analyses can identify the bottlenecks, guide the reliable process specifications and the most "informative" data collection, and improve production stability.

Vu Nguyen, Sunil Gupta, Santu Rana et al.

Bayesian optimization (BO) and its batch extensions are successful for optimizing expensive black-box functions. However, these traditional BO approaches are not yet ideal for optimizing less expensive functions when the computational cost of BO can dominate the cost of evaluating the blackbox function. Examples of these less expensive functions are cheap machine learning models, inexpensive physical experiment through simulators, and acquisition function optimization in Bayesian optimization. In this paper, we consider a batch BO setting for situations where function evaluations are less expensive. Our model is based on a new exploration strategy using geometric distance that provides an alternative way for exploration, selecting a point far from the observed locations. Using that intuition, we propose to use Sobol sequence to guide exploration that will get rid of running multiple global optimization steps as used in previous works. Based on the proposed distance exploration, we present an efficient batch BO approach. We demonstrate that our approach outperforms other baselines and global optimization methods when the function evaluations are less expensive.