Large Language Models (LLMs) have achieved remarkable success with their billion-level parameters, yet they incur high inference overheads. The emergence of activation sparsity in LLMs provides a natural approach to reduce this cost by involving only parts of the parameters for inference. Existing methods only focus on utilizing this naturally formed activation sparsity, overlooking the potential for further amplifying this inherent sparsity. In this paper, we hypothesize that LLMs can learn to be efficient by achieving more structured activation sparsity. To achieve this, we introduce a novel algorithm, Learn-To-be-Efficient (LTE), designed to train efficiency-aware LLMs to learn to activate fewer neurons and achieve a better trade-off between sparsity and performance. Furthermore, unlike SOTA MoEfication methods, which mainly focus on ReLU-based models, LTE can also be applied to LLMs like GPT and LLaMA with soft activation functions. We evaluate LTE on four models and eleven datasets. The experiments show that LTE achieves a better trade-off between sparsity and task performance. For instance, LTE with LLaMA provides a 1.83x-2.59x FLOPs speed-up on language generation tasks, outperforming the state-of-the-art methods.
相關內容
Large Language Models (LLMs) have rapidly become important tools in Biomedical and Health Informatics (BHI), enabling new ways to analyze data, treat patients, and conduct research. This bibliometric review aims to provide a panoramic view of how LLMs have been used in BHI by examining research articles and collaboration networks from 2022 to 2023. It further explores how LLMs can improve Natural Language Processing (NLP) applications in various BHI areas like medical diagnosis, patient engagement, electronic health record management, and personalized medicine. To do this, our bibliometric review identifies key trends, maps out research networks, and highlights major developments in this fast-moving field. Lastly, it discusses the ethical concerns and practical challenges of using LLMs in BHI, such as data privacy and reliable medical recommendations. Looking ahead, we consider how LLMs could further transform biomedical research as well as healthcare delivery and patient outcomes. This bibliometric review serves as a resource for stakeholders in healthcare, including researchers, clinicians, and policymakers, to understand the current state and future potential of LLMs in BHI.
Denoising Diffusion Probabilistic Models (DDPMs) have accomplished much in the realm of generative AI. Despite their high performance, there is room for improvement, especially in terms of sample fidelity by utilizing statistical properties that impose structural integrity, such as isotropy. Minimizing the mean squared error between the additive and predicted noise alone does not impose constraints on the predicted noise to be isotropic. Thus, we were motivated to utilize the isotropy of the additive noise as a constraint on the objective function to enhance the fidelity of DDPMs. Our approach is simple and can be applied to any DDPM variant. We validate our approach by presenting experiments conducted on four synthetic 2D datasets as well as on unconditional image generation. As demonstrated by the results, the incorporation of this constraint improves the fidelity metrics, Precision and Density for the 2D datasets as well as for the unconditional image generation.
Reinforcement Learning with Human Feedback (RLHF) has received significant attention for performing tasks without the need for costly manual reward design by aligning human preferences. It is crucial to consider diverse human feedback types and various learning methods in different environments. However, quantifying progress in RLHF with diverse feedback is challenging due to the lack of standardized annotation platforms and widely used unified benchmarks. To bridge this gap, we introduce Uni-RLHF, a comprehensive system implementation tailored for RLHF. It aims to provide a complete workflow from real human feedback, fostering progress in the development of practical problems. Uni-RLHF contains three packages: 1) a universal multi-feedback annotation platform, 2) large-scale crowdsourced feedback datasets, and 3) modular offline RLHF baseline implementations. Uni-RLHF develops a user-friendly annotation interface tailored to various feedback types, compatible with a wide range of mainstream RL environments. We then establish a systematic pipeline of crowdsourced annotations, resulting in large-scale annotated datasets comprising more than 15 million steps across 30+ popular tasks. Through extensive experiments, the results in the collected datasets demonstrate competitive performance compared to those from well-designed manual rewards. We evaluate various design choices and offer insights into their strengths and potential areas of improvement. We wish to build valuable open-source platforms, datasets, and baselines to facilitate the development of more robust and reliable RLHF solutions based on realistic human feedback. The website is available at //uni-rlhf.github.io/.
Multimodal Large Language Models (MLLMs) have showcased impressive skills in tasks related to visual understanding and reasoning. Yet, their widespread application faces obstacles due to the high computational demands during both the training and inference phases, restricting their use to a limited audience within the research and user communities. In this paper, we investigate the design aspects of Multimodal Small Language Models (MSLMs) and propose an efficient multimodal assistant named Mipha, which is designed to create synergy among various aspects: visual representation, language models, and optimization strategies. We show that without increasing the volume of training data, our Mipha-3B outperforms the state-of-the-art large MLLMs, especially LLaVA-1.5-13B, on multiple benchmarks. Through detailed discussion, we provide insights and guidelines for developing strong MSLMs that rival the capabilities of MLLMs. Our code is available at //github.com/zhuyiche/llava-phi.
Although Large Language Models (LLMs) have made significant progress in code generation, they still struggle with code generation tasks in specific scenarios. These scenarios usually necessitate the adaptation of LLMs to fulfill specific needs, but the limited training samples available in practice lead to poor code generation performance. Therefore, how to effectively adapt LLMs to new scenarios with few training samples is a major challenge for current code generation. In this paper, we propose a novel adaptation approach named SEED, which stands for Sample-Efficient adaptation with Error-Driven learning for code generation. SEED leverages the errors made by LLMs as learning opportunities, using error revision to overcome its own shortcomings, thus achieving efficient learning. Specifically, SEED involves identifying error code generated by LLMs, employing Self-revise for code revision, optimizing the model with revised code, and iteratively adapting the process for continuous improvement. Experimental results show that, compared to other mainstream fine-tuning approaches, SEED achieves superior performance with few training samples, showing an average relative improvement of 54.7% in Pass@1 on multiple code generation benchmarks. We also validate the effectiveness of Self-revise, which generates revised code that optimizes the model more efficiently compared to the code samples from datasets. Moreover, SEED consistently demonstrates strong performance across various LLMs, underscoring its generalizability.
Risk-limiting audits (RLAs) are techniques for verifying the outcomes of large elections. While they provide rigorous guarantees of correctness, widespread adoption has been impeded by both efficiency concerns and the fact they offer statistical, rather than absolute, conclusions. We attend to both of these difficulties, defining new families of audits that improve efficiency and offer qualitative advances in statistical power. Our new audits are enabled by revisiting the standard notion of a cast-vote record so that it can declare multiple possible mark interpretations rather than a single decision; this can reflect the presence of ambiguous marks, which appear regularly on hand-marked ballots. We show that this simple expedient can offer significant efficiency improvements with only minor changes to existing auditing infrastructure. We establish that these Bayesian comparison audits are indeed risk-limiting in the formal sense of (Fuller, Harrison, and Russell, 2022). We then define a new type of post-election audit we call a contested audit. These permit each candidate to provide a cast-vote record table advancing their own claim to victory. We prove that these audits offer remarkable sample efficiency, yielding control of risk with a constant number of samples (that is independent of margin). This is a first for an audit with provable soundness. These results are formulated in a game-based security model that specify quantitative soundness and completeness guarantees. Finally, we observe that these audits provide a direct means to handle contestation of election results affirmed by conventional RLAs.
Modern Large Language Models (LLMs) are capable of following long and complex instructions that enable a diverse amount of user tasks. However, despite Information Retrieval (IR) models using LLMs as the backbone of their architectures, nearly all of them still only take queries as input, with no instructions. For the handful of recent models that do take instructions, it's unclear how they use them. We introduce our dataset FollowIR, which contains a rigorous instruction evaluation benchmark as well as a training set for helping IR models learn to better follow real-world instructions. FollowIR builds off the long history of the TREC conferences: as TREC provides human annotators with instructions (also known as narratives) to determine document relevance, so should IR models be able to understand and decide relevance based on these detailed instructions. Our evaluation benchmark starts with three deeply judged TREC collections and alters the annotator instructions, re-annotating relevant documents. Through this process, we can measure how well IR models follow instructions, through a new pairwise evaluation framework. Our results indicate that existing retrieval models fail to correctly use instructions, using them for basic keywords and struggling to understand long-form information. However, we show that it is possible for IR models to learn to follow complex instructions: our new FollowIR-7B model has significant improvements (over 13%) after fine-tuning on our training set.
Collaborative Mobile Crowdsourcing (CMCS) allows platforms to recruit worker teams to collaboratively execute complex sensing tasks. The efficiency of such collaborations could be influenced by trust relationships among workers. To obtain the asymmetric trust values among all workers in the social network, the Trust Reinforcement Evaluation Framework (TREF) based on Graph Convolutional Neural Networks (GCNs) is proposed in this paper. The task completion effect is comprehensively calculated by considering the workers' ability benefits, distance benefits, and trust benefits in this paper. The worker recruitment problem is modeled as an Undirected Complete Recruitment Graph (UCRG), for which a specific Tabu Search Recruitment (TSR) algorithm solution is proposed. An optimal execution team is recruited for each task by the TSR algorithm, and the collaboration team for the task is obtained under the constraint of privacy loss. To enhance the efficiency of the recruitment algorithm on a large scale and scope, the Mini-Batch K-Means clustering algorithm and edge computing technology are introduced, enabling distributed worker recruitment. Lastly, extensive experiments conducted on five real datasets validate that the recruitment algorithm proposed in this paper outperforms other baselines. Additionally, TREF proposed herein surpasses the performance of state-of-the-art trust evaluation methods in the literature.
Existence constraints were defined in the Relational Data Model, but, unfortunately, are not provided by any Relational Database Management System, except for their NOT NULL particular case. Our (Elementary) Mathematical Data Model extended them to function products and introduced their dual non-existence constraints. MatBase, an intelligent data and knowledge base management system prototype based on both these data models, not only provides existence and non-existence constraints, but also automatically generates code for their enforcement. This paper presents and discusses the algorithms used by MatBase to enforce these types of constraints.
Knowledge Graph Embedding (KGE) aims to learn representations for entities and relations. Most KGE models have gained great success, especially on extrapolation scenarios. Specifically, given an unseen triple (h, r, t), a trained model can still correctly predict t from (h, r, ?), or h from (?, r, t), such extrapolation ability is impressive. However, most existing KGE works focus on the design of delicate triple modeling function, which mainly tells us how to measure the plausibility of observed triples, but offers limited explanation of why the methods can extrapolate to unseen data, and what are the important factors to help KGE extrapolate. Therefore in this work, we attempt to study the KGE extrapolation of two problems: 1. How does KGE extrapolate to unseen data? 2. How to design the KGE model with better extrapolation ability? For the problem 1, we first discuss the impact factors for extrapolation and from relation, entity and triple level respectively, propose three Semantic Evidences (SEs), which can be observed from train set and provide important semantic information for extrapolation. Then we verify the effectiveness of SEs through extensive experiments on several typical KGE methods. For the problem 2, to make better use of the three levels of SE, we propose a novel GNN-based KGE model, called Semantic Evidence aware Graph Neural Network (SE-GNN). In SE-GNN, each level of SE is modeled explicitly by the corresponding neighbor pattern, and merged sufficiently by the multi-layer aggregation, which contributes to obtaining more extrapolative knowledge representation. Finally, through extensive experiments on FB15k-237 and WN18RR datasets, we show that SE-GNN achieves state-of-the-art performance on Knowledge Graph Completion task and performs a better extrapolation ability.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191