PyTorchKR Papers

최신 논문

총 4,899편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2406.07067

TIM: Temporal Interaction Model in Notification System

Modern mobile applications heavily rely on the notification system to acquire daily active users and enhance user engagement. Being able to proactively reach users, the system has to decide when to send notifications to users. Although many researchers have studied optimizing the timing of sending notifications, they only utilized users' contextual features, without modeling users' behavior patterns. Additionally, these efforts only focus on individual notifications, and there is a lack of studies on optimizing the holistic timing of multiple notifications within a period. To bridge these gaps, we propose the Temporal Interaction Model (TIM), which models users' behavior patterns by estimating CTR in every time slot over a day in our short video application Kuaishou. TIM leverages long-term user historical interaction sequence features such as notification receipts, clicks, watch time and effective views, and employs a temporal attention unit (TAU) to extract user behavior patterns. Moreover, we provide an elegant strategy of holistic notifications send time control to improve user engagement while minimizing disruption. We evaluate the effectiveness of TIM through offline experiments and online A/B tests. The results indicate that TIM is a reliable tool for forecasting user behavior, leading to a remarkable enhancement in user engagement without causing undue disturbance.

arXiv:2406.05977

Weighted KL-Divergence for Document Ranking Model Refinement

Transformer-based retrieval and reranking models for text document search are often refined through knowledge distillation together with contrastive learning. A tight distribution matching between the teacher and student models can be hard as over-calibration may degrade training effectiveness when a teacher does not perform well. This paper contrastively reweights KL divergence terms to prioritize the alignment between a student and a teacher model for proper separation of positive and negative documents. This paper analyzes and evaluates the proposed loss function on the MS MARCO and BEIR datasets to demonstrate its effectiveness in improving the relevance of tested student models.

arXiv:2406.06371공식 코드 · Papers with Code

mHuBERT-147: A Compact Multilingual HuBERT Model

We present mHuBERT-147, the first general-purpose massively multilingual HuBERT speech representation model trained on 90K hours of clean, open-license data. To scale up the multi-iteration HuBERT approach, we use faiss-based clustering, achieving 5.2x faster label assignment than the original method. We also apply a new multilingual batching up-sampling strategy, leveraging both language and dataset diversity. After 3 training iterations, our compact 95M parameter mHuBERT-147 outperforms larger models trained on substantially more data. We rank second and first on the ML-SUPERB 10min and 1h leaderboards, with SOTA scores for 3 tasks. Across ASR/LID tasks, our model consistently surpasses XLS-R (300M params; 436K hours) and demonstrates strong competitiveness against the much larger MMS (1B params; 491K hours). Our findings indicate that mHuBERT-147 is a promising model for multilingual speech tasks, offering an unprecedented balance between high performance and parameter efficiency.

arXiv:2406.05615공식 코드 · Papers with Code

Video-Language Understanding: A Survey from Model Architecture, Model Training, and Data Perspectives

Humans use multiple senses to comprehend the environment. Vision and language are two of the most vital senses since they allow us to easily communicate our thoughts and perceive the world around us. There has been a lot of interest in creating video-language understanding systems with human-like senses since a video-language pair can mimic both our linguistic medium and visual environment with temporal dynamics. In this survey, we review the key tasks of these systems and highlight the associated challenges. Based on the challenges, we summarize their methods from model architecture, model training, and data perspectives. We also conduct performance comparison among the methods, and discuss promising directions for future research.

arXiv:2406.05741

Digital Business Model Analysis Using a Large Language Model

Digital transformation (DX) has recently become a pressing issue for many companies as the latest digital technologies, such as artificial intelligence and the Internet of Things, can be easily utilized. However, devising new business models is not easy for compa-nies, though they can improve their operations through digital technologies. Thus, business model design support methods are needed by people who lack digital tech-nology expertise. In contrast, large language models (LLMs) represented by ChatGPT and natural language processing utilizing LLMs have been developed revolutionarily. A business model design support system that utilizes these technologies has great potential. However, research on this area is scant. Accordingly, this study proposes an LLM-based method for comparing and analyzing similar companies from different business do-mains as a first step toward business model design support utilizing LLMs. This method can support idea generation in digital business model design.

arXiv:2406.05723공식 코드 · Papers with Code

Binarized Diffusion Model for Image Super-Resolution

Advanced diffusion models (DMs) perform impressively in image super-resolution (SR), but the high memory and computational costs hinder their deployment. Binarization, an ultra-compression algorithm, offers the potential for effectively accelerating DMs. Nonetheless, due to the model structure and the multi-step iterative attribute of DMs, existing binarization methods result in significant performance degradation. In this paper, we introduce a novel binarized diffusion model, BI-DiffSR, for image SR. First, for the model structure, we design a UNet architecture optimized for binarization. We propose the consistent-pixel-downsample (CP-Down) and consistent-pixel-upsample (CP-Up) to maintain dimension consistent and facilitate the full-precision information transfer. Meanwhile, we design the channel-shuffle-fusion (CS-Fusion) to enhance feature fusion in skip connection. Second, for the activation difference across timestep, we design the timestep-aware redistribution (TaR) and activation function (TaA). The TaR and TaA dynamically adjust the distribution of activations based on different timesteps, improving the flexibility and representation alability of the binarized module. Comprehensive experiments demonstrate that our BI-DiffSR outperforms existing binarization methods. Code is released at: https://github.com/zhengchen1999/BI-DiffSR.

arXiv:2406.06474

Towards a Personal Health Large Language Model

In health, most large language model (LLM) research has focused on clinical tasks. However, mobile and wearable devices, which are rarely integrated into such tasks, provide rich, longitudinal data for personal health monitoring. Here we present Personal Health Large Language Model (PH-LLM), fine-tuned from Gemini for understanding and reasoning over numerical time-series personal health data. We created and curated three datasets that test 1) production of personalized insights and recommendations from sleep patterns, physical activity, and physiological responses, 2) expert domain knowledge, and 3) prediction of self-reported sleep outcomes. For the first task we designed 857 case studies in collaboration with domain experts to assess real-world scenarios in sleep and fitness. Through comprehensive evaluation of domain-specific rubrics, we observed that Gemini Ultra 1.0 and PH-LLM are not statistically different from expert performance in fitness and, while experts remain superior for sleep, fine-tuning PH-LLM provided significant improvements in using relevant domain knowledge and personalizing information for sleep insights. We evaluated PH-LLM domain knowledge using multiple choice sleep medicine and fitness examinations. PH-LLM achieved 79% on sleep and 88% on fitness, exceeding average scores from a sample of human experts. Finally, we trained PH-LLM to predict self-reported sleep quality outcomes from textual and multimodal encoding representations of wearable data, and demonstrate that multimodal encoding is required to match performance of specialized discriminative models. Although further development and evaluation are necessary in the safety-critical personal health domain, these results demonstrate both the broad knowledge and capabilities of Gemini models and the benefit of contextualizing physiological data for personal health applications as done with PH-LLM.

arXiv:2406.04854

Uncertainty Aware Learning for Language Model Alignment

As instruction-tuned large language models (LLMs) evolve, aligning pretrained foundation models presents increasing challenges. Existing alignment strategies, which typically leverage diverse and high-quality data sources, often overlook the intrinsic uncertainty of tasks, learning all data samples equally. This may lead to suboptimal data efficiency and model performance. In response, we propose uncertainty-aware learning (UAL) to improve the model alignment of different task scenarios, by introducing the sample uncertainty (elicited from more capable LLMs). We implement UAL in a simple fashion -- adaptively setting the label smoothing value of training according to the uncertainty of individual samples. Analysis shows that our UAL indeed facilitates better token clustering in the feature space, validating our hypothesis. Extensive experiments on widely used benchmarks demonstrate that our UAL significantly and consistently outperforms standard supervised fine-tuning. Notably, LLMs aligned in a mixed scenario have achieved an average improvement of 10.62\% on high-entropy tasks (i.e., AlpacaEval leaderboard), and 1.81\% on complex low-entropy tasks (i.e., MetaMath and GSM8K).

arXiv:2406.04638

Large Language Model-guided Document Selection

Large Language Model (LLM) pre-training exhausts an ever growing compute budget, yet recent research has demonstrated that careful document selection enables comparable model quality with only a fraction of the FLOPs. Inspired by efforts suggesting that domain-specific training document selection is in fact an interpretable process [Gunasekar et al., 2023], as well as research showing that instruction-finetuned LLMs are adept zero-shot data labelers [Gilardi et al.,2023], we explore a promising direction for scalable general-domain document selection; employing a prompted LLM as a document grader, we distill quality labels into a classifier model, which is applied at scale to a large, and already heavily-filtered, web-crawl-derived corpus autonomously. Following the guidance of this classifier, we drop 75% of the corpus and train LLMs on the remaining data. Results across multiple benchmarks show that: 1. Filtering allows us to quality-match a model trained on the full corpus across diverse benchmarks with at most 70% of the FLOPs, 2. More capable LLM labelers and classifier models lead to better results that are less sensitive to the labeler's prompt, 3. In-context learning helps to boost the performance of less-capable labeling models. In all cases we use open-source datasets, models, recipes, and evaluation frameworks, so that results can be reproduced by the community.

arXiv:2406.04836공식 코드 · Papers with Code

Revisiting Catastrophic Forgetting in Large Language Model Tuning

Catastrophic Forgetting (CF) means models forgetting previously acquired knowledge when learning new data. It compromises the effectiveness of large language models (LLMs) during fine-tuning, yet the underlying causes have not been thoroughly investigated. This paper takes the first step to reveal the direct link between the flatness of the model loss landscape and the extent of CF in the field of LLMs. Based on this, we introduce the sharpness-aware minimization to mitigate CF by flattening the loss landscape. Experiments on three widely-used fine-tuning datasets, spanning different model scales, demonstrate the effectiveness of our method in alleviating CF. Analyses show that we nicely complement the existing anti-forgetting strategies, further enhancing the resistance of LLMs to CF.

arXiv:2406.04809

A Survey of Fragile Model Watermarking

Model fragile watermarking, inspired by both the field of adversarial attacks on neural networks and traditional multimedia fragile watermarking, has gradually emerged as a potent tool for detecting tampering, and has witnessed rapid development in recent years. Unlike robust watermarks, which are widely used for identifying model copyrights, fragile watermarks for models are designed to identify whether models have been subjected to unexpected alterations such as backdoors, poisoning, compression, among others. These alterations can pose unknown risks to model users, such as misidentifying stop signs as speed limit signs in classic autonomous driving scenarios. This paper provides an overview of the relevant work in the field of model fragile watermarking since its inception, categorizing them and revealing the developmental trajectory of the field, thus offering a comprehensive survey for future endeavors in model fragile watermarking.

논문 ID: document/1048739

Evolving Efficient CNN Based Model for Image Classification

Evolutionary algorithms, rooted in Darwin's theorem, can be considered as a base for implementing deep/machine learning models. This approach can noticeably increase the accuracy in most cases as shown in this paper. This experiment aims to evaluate the performance of two evolutionary algorithms, an evolutionary neural network (ENN) and an evolutionary CNN-based algorithm with mutation and crossover (ECNNB), on the Fashion-MNIST, MNIST, and EMNIST_Digits datasets. The performance of the ENN algorithm is examined for 10, 50, and 100 generations, with 50 generations being used due to computational limitations. The results show that the accuracy of the model improves as the number of generations increases. However, the ECNNB model consistently outperforms the ENN model on all three datasets, with an average accuracy of 92.58% on Fashion-MNIST, 99.32% on MNIST, and 99.50% on EMNIST_Digits, compared to 88.54 %, 98.05 %, and 98.95 %, respectively, for the ENN model. The performance of both models is compared with other state-of-the-art models in the literature. These results highlight the significance of well-designed models in achieving high accuracy in machine learning tasks.

arXiv:2406.03562공식 코드 · Papers with Code

Neural empirical interpolation method for nonlinear model reduction

In this paper, we introduce the neural empirical interpolation method (NEIM), a neural network-based alternative to the discrete empirical interpolation method for reducing the time complexity of computing the nonlinear term in a reduced order model (ROM) for a parameterized nonlinear partial differential equation. NEIM is a greedy algorithm which accomplishes this reduction by approximating an affine decomposition of the nonlinear term of the ROM, where the vector terms of the expansion are given by neural networks depending on the ROM solution, and the coefficients are given by an interpolation of some "optimal" coefficients. Because NEIM is based on a greedy strategy, we are able to provide a basic error analysis to investigate its performance. NEIM has the advantages of being easy to implement in models with automatic differentiation, of being a nonlinear projection of the ROM nonlinearity, of being efficient for both nonlocal and local nonlinearities, and of relying solely on data and not the explicit form of the ROM nonlinearity. We demonstrate the effectiveness of the methodology on solution-dependent and solution-independent nonlinearities, a nonlinear elliptic problem, and a nonlinear parabolic model of liquid crystals. Code availability: https://github.com/maxhirsch/NEIM

arXiv:2406.04333공식 코드 · Papers with Code

BitsFusion: 1.99 bits Weight Quantization of Diffusion Model

Diffusion-based image generation models have achieved great success in recent years by showing the capability of synthesizing high-quality content. However, these models contain a huge number of parameters, resulting in a significantly large model size. Saving and transferring them is a major bottleneck for various applications, especially those running on resource-constrained devices. In this work, we develop a novel weight quantization method that quantizes the UNet from Stable Diffusion v1.5 to 1.99 bits, achieving a model with 7.9X smaller size while exhibiting even better generation quality than the original one. Our approach includes several novel techniques, such as assigning optimal bits to each layer, initializing the quantized model for better performance, and improving the training strategy to dramatically reduce quantization error. Furthermore, we extensively evaluate our quantized model across various benchmark datasets and through human evaluation to demonstrate its superior generation quality.

arXiv:2406.04324공식 코드 · Papers with Code

SF-V: Single Forward Video Generation Model

Diffusion-based video generation models have demonstrated remarkable success in obtaining high-fidelity videos through the iterative denoising process. However, these models require multiple denoising steps during sampling, resulting in high computational costs. In this work, we propose a novel approach to obtain single-step video generation models by leveraging adversarial training to fine-tune pre-trained video diffusion models. We show that, through the adversarial training, the multi-steps video diffusion model, i.e., Stable Video Diffusion (SVD), can be trained to perform single forward pass to synthesize high-quality videos, capturing both temporal and spatial dependencies in the video data. Extensive experiments demonstrate that our method achieves competitive generation quality of synthesized videos with significantly reduced computational overhead for the denoising process (i.e., around $23\times$ speedup compared with SVD and $6\times$ speedup compared with existing works, with even better generation quality), paving the way for real-time video synthesis and editing. More visualization results are made publicly available at https://snap-research.github.io/SF-V.

arXiv:2406.03507

Robust Prediction Model for Multidimensional and Unbalanced Datasets

Data Mining is a promising field and is applied in multiple domains for its predictive capabilities. Data in the real world cannot be readily used for data mining as it suffers from the problems of multidimensionality, unbalance and missing values. It is difficult to use its predictive capabilities by novice users. It is difficult for a beginner to find the relevant set of attributes from a large pool of data available. The paper presents a Robust Prediction Model that finds a relevant set of attributes; resolves the problems of unbalanced and multidimensional real-life datasets and helps in finding patterns for informed decision making. Model is tested upon five different datasets in the domain of Health Sector, Education, Business and Fraud Detection. The results showcase the robust behaviour of the model and its applicability in various domains.

arXiv:2406.03733

Credit Card Fraud Detection Using Advanced Transformer Model

With the proliferation of various online and mobile payment systems, credit card fraud has emerged as a significant threat to financial security. This study focuses on innovative applications of the latest Transformer models for more robust and precise fraud detection. To ensure the reliability of the data, we meticulously processed the data sources, balancing the dataset to address the issue of data sparsity significantly. We also selected highly correlated vectors to strengthen the training process.To guarantee the reliability and practicality of the new Transformer model, we conducted performance comparisons with several widely adopted models, including Support Vector Machine (SVM), Random Forest, Neural Network, and Logistic Regression. We rigorously compared these models using metrics such as Precision, Recall, and F1 Score. Through these detailed analyses and comparisons, we present to the readers a highly efficient and powerful anti-fraud mechanism with promising prospects. The results demonstrate that the Transformer model not only excels in traditional applications but also shows great potential in niche areas like fraud detection, offering a substantial advancement in the field.

arXiv:2406.03689공식 코드 · Papers with Code

Evaluating the World Model Implicit in a Generative Model

Recent work suggests that large language models may implicitly learn world models. How should we assess this possibility? We formalize this question for the case where the underlying reality is governed by a deterministic finite automaton. This includes problems as diverse as simple logical reasoning, geographic navigation, game-playing, and chemistry. We propose new evaluation metrics for world model recovery inspired by the classic Myhill-Nerode theorem from language theory. We illustrate their utility in three domains: game playing, logic puzzles, and navigation. In all domains, the generative models we consider do well on existing diagnostics for assessing world models, but our evaluation metrics reveal their world models to be far less coherent than they appear. Such incoherence creates fragility: using a generative model to solve related but subtly different tasks can lead to failures. Building generative models that meaningfully capture the underlying logic of the domains they model would be immensely valuable; our results suggest new ways to assess how close a given model is to that goal.

arXiv:2406.00164

DYNA: Disease-Specific Language Model for Variant Pathogenicity

Clinical variant classification of pathogenic versus benign genetic variants remains a challenge in clinical genetics. Recently, the proposition of genomic foundation models has improved the generic variant effect prediction (VEP) accuracy via weakly-supervised or unsupervised training. However, these VEPs are not disease-specific, limiting their adaptation at the point of care. To address this problem, we propose DYNA: Disease-specificity fine-tuning via a Siamese neural network broadly applicable to all genomic foundation models for more effective variant effect predictions in disease-specific contexts. We evaluate DYNA in two distinct disease-relevant tasks. For coding VEPs, we focus on various cardiovascular diseases, where gene-disease relationships of loss-of-function vs. gain-of-function dictate disease-specific VEP. For non-coding VEPs, we apply DYNA to an essential post-transcriptional regulatory axis of RNA splicing, the most common non-coding pathogenic mechanism in established clinical VEP guidelines. In both cases, DYNA fine-tunes various pre-trained genomic foundation models on small, rare variant sets. The DYNA fine-tuned models show superior performance in the held-out rare variant testing set and are further replicated in large, clinically-relevant variant annotations in ClinVAR. Thus, DYNA offers a potent disease-specific variant effect prediction method, excelling in intra-gene generalization and generalization to unseen genetic variants, making it particularly valuable for disease associations and clinical applicability.

arXiv:2406.02887

USM RNN-T model weights binarization

Large-scale universal speech models (USM) are already used in production. However, as the model size grows, the serving cost grows too. Serving cost of large models is dominated by model size that is why model size reduction is an important research topic. In this work we are focused on model size reduction using weights only quantization. We present the weights binarization of USM Recurrent Neural Network Transducer (RNN-T) and show that its model size can be reduced by 15.9x times at cost of word error rate (WER) increase by only 1.9% in comparison to the float32 model. It makes it attractive for practical applications.