The boom in Large Language Models (LLMs) like GPT-4 and ChatGPT has marked a significant advancement in artificial intelligence. These models are becoming increasingly complex and powerful to train and serve. This growth in capabilities comes with a substantial increase in computational requirements, both in terms of hardware resources and energy consumption. The goal of this paper is to showcase how hardware and software co-design can come together and allow us to create customized hardware systems for specific LLM workloads. We propose a simulation workflow that allows us to combine model parallelism techniques with a multi-accelerator simulation framework for efficiency metrics. We focus on inference workloads and report power, cycle, and latency metrics upon performing a design space exploration search over multiple software and hardware configurations.

Individuals with suspected rare genetic disorders often undergo multiple clinical evaluations, imaging studies, laboratory tests and genetic tests, to find a possible answer over a prolonged period of multiple years. Addressing this diagnostic odyssey thus have substantial clinical, psychosocial, and economic benefits. Many rare genetic diseases have distinctive facial features, which can be used by artificial intelligence algorithms to facilitate clinical diagnosis, in prioritizing candidate diseases to be further examined by lab tests or genetic assays, or in helping the phenotype-driven reinterpretation of genome/exome sequencing data. However, existing methods using frontal facial photo were built on conventional Convolutional Neural Networks (CNNs), rely exclusively on facial images, and cannot capture non-facial phenotypic traits and demographic information essential for guiding accurate diagnoses. Here we introduce GestaltMML, a multimodal machine learning (MML) approach solely based on the Transformer architecture. It integrates the facial images, demographic information (age, sex, ethnicity), and clinical notes of patients to improve prediction accuracy. Furthermore, we also introduce GestaltGPT, a GPT-based methodology with few-short learning capacities that exclusively harnesses textual inputs using a range of large language models (LLMs) including Llama 2, GPT-J and Falcon. We evaluated these methods on a diverse range of datasets, including 449 diseases from the GestaltMatcher Database, several in-house datasets on Beckwith-Wiedemann syndrome, Sotos syndrome, NAA10-related syndrome (neurodevelopmental syndrome) and others. Our results suggest that GestaltMML/GestaltGPT effectively incorporate multiple modalities of data, greatly narrow down candidate genetic diagnosis of rare diseases, and may facilitate the reinterpretation of genome/exome sequencing data.

ChatGPT's proficiency in handling modern standard languages suggests potential for its use in understanding ancient Chinese. This project explores ChatGPT's capabilities on ancient Chinese via two tasks: translating ancient Chinese to modern Chinese and recognizing ancient Chinese names. A comparison of ChatGPT's output with human translations serves to evaluate its comprehension of ancient Chinese. The findings indicate that: (1.)the proficiency of ancient Chinese by ChatGPT is yet to reach a satisfactory level; (2.) ChatGPT performs the best on ancient-to-modern translation when feeding with three context sentences. To help reproduce our work, we display the python code snippets used in this study.

In the rapidly evolving landscape of artificial intelligence (AI), generative large language models (LLMs) stand at the forefront, revolutionizing how we interact with our data. However, the computational intensity and memory consumption of deploying these models present substantial challenges in terms of serving efficiency, particularly in scenarios demanding low latency and high throughput. This survey addresses the imperative need for efficient LLM serving methodologies from a machine learning system (MLSys) research perspective, standing at the crux of advanced AI innovations and practical system optimizations. We provide in-depth analysis, covering a spectrum of solutions, ranging from cutting-edge algorithmic modifications to groundbreaking changes in system designs. The survey aims to provide a comprehensive understanding of the current state and future directions in efficient LLM serving, offering valuable insights for researchers and practitioners in overcoming the barriers of effective LLM deployment, thereby reshaping the future of AI.

Neural Networks can be efficiently compressed through pruning, significantly reducing storage and computational demands while maintaining predictive performance. Simple yet effective methods like Iterative Magnitude Pruning (IMP, Han et al., 2015) remove less important parameters and require a costly retraining procedure to recover performance after pruning. However, with the rise of Large Language Models (LLMs), full retraining has become infeasible due to memory and compute constraints. In this study, we challenge the practice of retraining all parameters by demonstrating that updating only a small subset of highly expressive parameters is often sufficient to recover or even improve performance compared to full retraining. Surprisingly, retraining as little as 0.27%-0.35% of the parameters of GPT-architectures (OPT-2.7B/6.7B/13B/30B) achieves comparable performance to One Shot IMP across various sparsity levels. Our method, Parameter-Efficient Retraining after Pruning (PERP), drastically reduces compute and memory demands, enabling pruning and retraining of up to 30 billion parameter models on a single NVIDIA A100 GPU within minutes. Despite magnitude pruning being considered as unsuited for pruning LLMs, our findings show that PERP positions it as a strong contender against state-of-the-art retraining-free approaches such as Wanda (Sun et al., 2023) and SparseGPT (Frantar & Alistarh, 2023), opening up a promising alternative to avoiding retraining.

The advances of Large Language Models (LLMs) are expanding their utility in both academic research and practical applications. Recent social science research has explored the use of these "black-box" LLM agents for simulating complex social systems and potentially substituting human subjects in experiments. Our study delves into this emerging domain, investigating the extent to which LLMs exhibit key social interaction principles, such as social learning, social preference, and cooperative behavior, in their interactions with humans and other agents. We develop a novel framework for our study, wherein classical laboratory experiments involving human subjects are adapted to use LLM agents. This approach involves step-by-step reasoning that mirrors human cognitive processes and zero-shot learning to assess the innate preferences of LLMs. Our analysis of LLM agents' behavior includes both the primary effects and an in-depth examination of the underlying mechanisms. Focusing on GPT-4, the state-of-the-art LLM, our analyses suggest that LLM agents appear to exhibit a range of human-like social behaviors such as distributional and reciprocity preferences, responsiveness to group identity cues, engagement in indirect reciprocity, and social learning capabilities. However, our analysis also reveals notable differences: LLMs demonstrate a pronounced fairness preference, weaker positive reciprocity, and a more calculating approach in social learning compared to humans. These insights indicate that while LLMs hold great promise for applications in social science research, such as in laboratory experiments and agent-based modeling, the subtle behavioral differences between LLM agents and humans warrant further investigation. Careful examination and development of protocols in evaluating the social behaviors of LLMs are necessary before directly applying these models to emulate human behavior.

Multi-hop question answering (MQA) is one of the challenging tasks to evaluate machine's comprehension and reasoning abilities, where large language models (LLMs) have widely achieved the human-comparable performance. Due to the dynamics of knowledge facts in real world, knowledge editing has been explored to update model with the up-to-date facts while avoiding expensive re-training or fine-tuning. Starting from the edited fact, the updated model needs to provide cascading changes in the chain of MQA. The previous art simply adopts a mix-up prompt to instruct LLMs conducting multiple reasoning tasks sequentially, including question decomposition, answer generation, and conflict checking via comparing with edited facts. However, the coupling of these functionally-diverse reasoning tasks inhibits LLMs' advantages in comprehending and answering questions while disturbing them with the unskilled task of conflict checking. We thus propose a framework, Programmable knowledge editing for Multi-hop Question Answering (PokeMQA), to decouple the jobs. Specifically, we prompt LLMs to decompose knowledge-augmented multi-hop question, while interacting with a detached trainable scope detector to modulate LLMs behavior depending on external conflict signal. The experiments on three LLM backbones and two benchmark datasets validate our superiority in knowledge editing of MQA, outperforming all competitors by a large margin in almost all settings and consistently producing reliable reasoning process.

Lowering the memory requirement in full-parameter training on large models has become a hot research area. MeZO fine-tunes the large language models (LLMs) by just forward passes in a zeroth-order SGD optimizer (ZO-SGD), demonstrating excellent performance with the same GPU memory usage as inference. However, the simulated perturbation stochastic approximation for gradient estimate in MeZO leads to severe oscillations and incurs a substantial time overhead. Moreover, without momentum regularization, MeZO shows severe over-fitting problems. Lastly, the perturbation-irrelevant momentum on ZO-SGD does not improve the convergence rate. This study proposes ZO-AdaMU to resolve the above problems by adapting the simulated perturbation with momentum in its stochastic approximation. Unlike existing adaptive momentum methods, we relocate momentum on simulated perturbation in stochastic gradient approximation. Our convergence analysis and experiments prove this is a better way to improve convergence stability and rate in ZO-SGD. Extensive experiments demonstrate that ZO-AdaMU yields better generalization for LLMs fine-tuning across various NLP tasks than MeZO and its momentum variants.

Lack of diverse perspectives causes neutrality bias in Wikipedia content leading to millions of worldwide readers getting exposed by potentially inaccurate information. Hence, neutrality bias detection and mitigation is a critical problem. Although previous studies have proposed effective solutions for English, no work exists for Indian languages. First, we contribute two large datasets, mWikiBias and mWNC, covering 8 languages, for the bias detection and mitigation tasks respectively. Next, we investigate the effectiveness of popular multilingual Transformer-based models for the two tasks by modeling detection as a binary classification problem and mitigation as a style transfer problem. We make the code and data publicly available.

User modeling (UM) aims to discover patterns or learn representations from user data about the characteristics of a specific user, such as profile, preference, and personality. The user models enable personalization and suspiciousness detection in many online applications such as recommendation, education, and healthcare. Two common types of user data are text and graph, as the data usually contain a large amount of user-generated content (UGC) and online interactions. The research of text and graph mining is developing rapidly, contributing many notable solutions in the past two decades. Recently, large language models (LLMs) have shown superior performance on generating, understanding, and even reasoning over text data. The approaches of user modeling have been equipped with LLMs and soon become outstanding. This article summarizes existing research about how and why LLMs are great tools of modeling and understanding UGC. Then it reviews a few categories of large language models for user modeling (LLM-UM) approaches that integrate the LLMs with text and graph-based methods in different ways. Then it introduces specific LLM-UM techniques for a variety of UM applications. Finally, it presents remaining challenges and future directions in the LLM-UM research. We maintain the reading list at: https://github.com/TamSiuhin/LLM-UM-Reading