In this paper, we propose a novel parameter and computation efficient tuning method for Multi-modal Large Language Models (MLLMs), termed Efficient Attention Skipping (EAS). Concretely, we first reveal that multi-head attentions (MHAs), the main computational overhead of MLLMs, are often redundant to downstream tasks. Based on this observation, EAS evaluates the attention redundancy and skips the less important MHAs to speed up inference. Besides, we also propose a novel propagation-of-information adapter (PIA) to serve the attention skipping of EAS and keep parameter efficiency, which can be further re-parameterized into feed-forward networks (FFNs) for zero-extra latency. To validate EAS, we apply it to a recently proposed MLLM called LaVIN and a classic VL pre-trained model called METER, and conduct extensive experiments on a set of benchmarks. The experiments show that EAS not only retains high performance and parameter efficiency, but also greatly speeds up inference speed. For instance, LaVIN-EAS can obtain 89.98\% accuracy on ScineceQA while speeding up inference by 2.2 times to LaVIN
相關內容
We present SCULPT, a novel 3D generative model for clothed and textured 3D meshes of humans. Specifically, we devise a deep neural network that learns to represent the geometry and appearance distribution of clothed human bodies. Training such a model is challenging, as datasets of textured 3D meshes for humans are limited in size and accessibility. Our key observation is that there exist medium-sized 3D scan datasets like CAPE, as well as large-scale 2D image datasets of clothed humans and multiple appearances can be mapped to a single geometry. To effectively learn from the two data modalities, we propose an unpaired learning procedure for pose-dependent clothed and textured human meshes. Specifically, we learn a pose-dependent geometry space from 3D scan data. We represent this as per vertex displacements w.r.t. the SMPL model. Next, we train a geometry conditioned texture generator in an unsupervised way using the 2D image data. We use intermediate activations of the learned geometry model to condition our texture generator. To alleviate entanglement between pose and clothing type, and pose and clothing appearance, we condition both the texture and geometry generators with attribute labels such as clothing types for the geometry, and clothing colors for the texture generator. We automatically generated these conditioning labels for the 2D images based on the visual question answering model BLIP and CLIP. We validate our method on the SCULPT dataset, and compare to state-of-the-art 3D generative models for clothed human bodies. Our code and data can be found at //sculpt.is.tue.mpg.de.
In this paper, we investigate unsourced random access for massive machine-type communications (mMTC) in the sixth-generation (6G) wireless networks. Firstly, we establish a high-efficiency uncoupled framework for massive unsourced random access without extra parity check bits. Then, we design a low-complexity Bayesian joint decoding algorithm, including codeword detection and stitching. In particular, we present a Bayesian codeword detection approach by exploiting Bayes-optimal divergence-free orthogonal approximate message passing in the case of unknown priors. The output long-term channel statistic information is well leveraged to stitch codewords for recovering the original message. Thus, the spectral efficiency is improved by avoiding the use of parity bits. Moreover, we analyze the performance of the proposed Bayesian joint decoding-based massive uncoupled unsourced random access scheme in terms of computational complexity and error probability of decoding. Furthermore, by asymptotic analysis, we obtain some useful insights for the design of massive unsourced random access. Finally, extensive simulation results confirm the effectiveness of the proposed scheme in 6G wireless networks.
In this paper, we study a remote monitoring system where a receiver observes a remote binary Markov source and decides whether to sample and fetch the source's state over a randomly delayed channel. Due to transmission delay, the observation of the source is imperfect, resulting in the uncertainty of the source's state at the receiver. We thus use uncertainty of information as the metric to characterize the performance of the system. Measured by Shannon's entropy, uncertainty of information reflects how much we do not know about the latest source's state in the absence of new information. The current research for uncertainty of information idealizes the transmission delay as one time slot, but not under random delay. Moreover, uncertainty of information varies with the latest observation of the source's state, making it different from other age of information related functions. Motivated by the above reasons, we formulate a uncertainty of information minimization problem under random delay. Typically, such a problem which takes actions based on the imperfect observations can be modeled as a partially observed Markov decision process. By introducing belief state, we transform this process into a semi-Markov decision process. To solve this problem, we first provide an optimal sampling policy employing a two layered bisection relative value iteration algorithm. Furthermore, we propose a sub-optimal index policy with low complexity based on the special properties of belief state. Numerical simulations illustrate that both of the proposed sampling policies outperforms two other benchmarks. Moreover, the performance of the sub-optimal policy approaches to that of the optimal policy, particularly under large delay.
In this paper, we propose a novel model for a malware classification system based on Application Programming Interface (API) calls and opcodes, to improve classification accuracy. This system uses a novel design of combined Convolutional Neural Network and Long Short-Term Memory. We extract opcode sequences and API Calls from Windows malware samples for classification. We transform these features into N-grams (N = 2, 3, and 10)-gram sequences. Our experiments on a dataset of 9,749,57 samples produce high accuracy of 99.91% using the 8-gram sequences. Our method significantly improves the malware classification performance when using a wide range of recent deep learning architectures, leading to state-of-the-art performance. In particular, we experiment with ConvNeXt-T, ConvNeXt-S, RegNetY-4GF, RegNetY-8GF, RegNetY-12GF, EfficientNetV2, Sequencer2D-L, Swin-T, ViT-G/14, ViT-Ti, ViT-S, VIT-B, VIT-L, and MaxViT-B. Among these architectures, Swin-T and Sequencer2D-L architectures achieved high accuracies of 99.82% and 99.70%, respectively, comparable to our CNN-LSTM architecture although not surpassing it.
In this paper, we present a novel approach for text independent phone-to-audio alignment based on phoneme recognition, representation learning and knowledge transfer. Our method leverages a self-supervised model (wav2vec2) fine-tuned for phoneme recognition using a Connectionist Temporal Classification (CTC) loss, a dimension reduction model and a frame-level phoneme classifier trained thanks to forced-alignment labels (using Montreal Forced Aligner) to produce multi-lingual phonetic representations, thus requiring minimal additional training. We evaluate our model using synthetic native data from the TIMIT dataset and the SCRIBE dataset for American and British English, respectively. Our proposed model outperforms the state-of-the-art (charsiu) in statistical metrics and has applications in language learning and speech processing systems. We leave experiments on other languages for future work but the design of the system makes it easily adaptable to other languages.
In this paper, we introduce an alternative approach to enhancing Multi-Agent Reinforcement Learning (MARL) through the integration of domain knowledge and attention-based policy mechanisms. Our methodology focuses on the incorporation of domain-specific expertise into the learning process, which simplifies the development of collaborative behaviors. This approach aims to reduce the complexity and learning overhead typically associated with MARL by enabling agents to concentrate on essential aspects of complex tasks, thus optimizing the learning curve. The utilization of attention mechanisms plays a key role in our model. It allows for the effective processing of dynamic context data and nuanced agent interactions, leading to more refined decision-making. Applied in standard MARL scenarios, such as the Stanford Intelligent Systems Laboratory (SISL) Pursuit and Multi-Particle Environments (MPE) Simple Spread, our method has been shown to improve both learning efficiency and the effectiveness of collaborative behaviors. The results indicate that our attention-based approach can be a viable approach for improving the efficiency of MARL training process, integrating domain-specific knowledge at the action level.
This paper presents an efficient finite element iterative method for solving a nonuniform size-modified Poisson-Nernst-Planck ion channel (SMPNPIC) model, along with a SMPNPIC program package that works for an ion channel protein with a three-dimensional crystallographic structure and an ionic solvent with multiple ionic species. In particular, the SMPNPIC model is constructed and then reformulated by novel mathematical techniques so that each iteration of the method only involves linear boundary value problems and nonlinear algebraic systems, circumventing the numerical difficulties caused by the strong nonlinearities, strong asymmetries, and strong differential equation coupling of the SMPNPIC model. To further improve the method's efficiency, an efficient modified Newton iterative method is adapted to the numerical solution of each related nonlinear algebraic system. Numerical results for a voltage-dependent anion channel (VDAC) and a mixture solution of four ionic species demonstrate the method's convergence, the package's high performance, and the importance of considering nonuniform ion size effects. They also partially validate the SMPNPIC model by the anion selectivity property of VDAC.
This paper presents an exhaustive quantitative and qualitative evaluation of Large Language Models (LLMs) for Knowledge Graph (KG) construction and reasoning. We employ eight distinct datasets that encompass aspects including entity, relation and event extraction, link prediction, and question answering. Empirically, our findings suggest that GPT-4 outperforms ChatGPT in the majority of tasks and even surpasses fine-tuned models in certain reasoning and question-answering datasets. Moreover, our investigation extends to the potential generalization ability of LLMs for information extraction, which culminates in the presentation of the Virtual Knowledge Extraction task and the development of the VINE dataset. Drawing on these empirical findings, we further propose AutoKG, a multi-agent-based approach employing LLMs for KG construction and reasoning, which aims to chart the future of this field and offer exciting opportunities for advancement. We anticipate that our research can provide invaluable insights for future undertakings of KG\footnote{Code and datasets will be available in //github.com/zjunlp/AutoKG.
This paper presents a comprehensive and practical guide for practitioners and end-users working with Large Language Models (LLMs) in their downstream natural language processing (NLP) tasks. We provide discussions and insights into the usage of LLMs from the perspectives of models, data, and downstream tasks. Firstly, we offer an introduction and brief summary of current GPT- and BERT-style LLMs. Then, we discuss the influence of pre-training data, training data, and test data. Most importantly, we provide a detailed discussion about the use and non-use cases of large language models for various natural language processing tasks, such as knowledge-intensive tasks, traditional natural language understanding tasks, natural language generation tasks, emergent abilities, and considerations for specific tasks.We present various use cases and non-use cases to illustrate the practical applications and limitations of LLMs in real-world scenarios. We also try to understand the importance of data and the specific challenges associated with each NLP task. Furthermore, we explore the impact of spurious biases on LLMs and delve into other essential considerations, such as efficiency, cost, and latency, to ensure a comprehensive understanding of deploying LLMs in practice. This comprehensive guide aims to provide researchers and practitioners with valuable insights and best practices for working with LLMs, thereby enabling the successful implementation of these models in a wide range of NLP tasks. A curated list of practical guide resources of LLMs, regularly updated, can be found at \url{//github.com/Mooler0410/LLMsPracticalGuide}.
In this paper, we propose a novel Feature Decomposition and Reconstruction Learning (FDRL) method for effective facial expression recognition. We view the expression information as the combination of the shared information (expression similarities) across different expressions and the unique information (expression-specific variations) for each expression. More specifically, FDRL mainly consists of two crucial networks: a Feature Decomposition Network (FDN) and a Feature Reconstruction Network (FRN). In particular, FDN first decomposes the basic features extracted from a backbone network into a set of facial action-aware latent features to model expression similarities. Then, FRN captures the intra-feature and inter-feature relationships for latent features to characterize expression-specific variations, and reconstructs the expression feature. To this end, two modules including an intra-feature relation modeling module and an inter-feature relation modeling module are developed in FRN. Experimental results on both the in-the-lab databases (including CK+, MMI, and Oulu-CASIA) and the in-the-wild databases (including RAF-DB and SFEW) show that the proposed FDRL method consistently achieves higher recognition accuracy than several state-of-the-art methods. This clearly highlights the benefit of feature decomposition and reconstruction for classifying expressions.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191