PyTorchKR Papers

최신 논문

총 398편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2411.10086공식 코드 · Papers with Code

CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation

Open-vocabulary semantic segmentation aims to assign semantic labels to each pixel without relying on a predefined set of categories. Contrastive Language-Image Pre-training (CLIP) demonstrates outstanding zero-shot classification capabilities but struggles with the pixel-wise segmentation task as the captured inter-patch correlations correspond to no specific visual concepts. Despite previous CLIP-based works improving inter-patch correlations by self-self attention, they still face the inherent limitation that image patches tend to have high similarity to outlier ones. In this work, we introduce CorrCLIP, a training-free approach for open-vocabulary semantic segmentation, which reconstructs significantly coherent inter-patch correlations utilizing foundation models. Specifically, it employs the Segment Anything Model (SAM) to define the scope of patch interactions, ensuring that patches interact only with semantically similar ones. Furthermore, CorrCLIP obtains an understanding of an image's semantic layout via self-supervised models to determine concrete similarity values between image patches, which addresses the similarity irregularity problem caused by the aforementioned restricted patch interaction regime. Finally, CorrCLIP reuses the region masks produced by SAM to update the segmentation map. As a training-free method, CorrCLIP achieves a notable improvement across eight challenging benchmarks regarding the averaged mean Intersection over Union, boosting it from 44.4% to 51.0%.

arXiv:2411.08687

Measuring similarity between embedding spaces using induced neighborhood graphs

Deep Learning techniques have excelled at generating embedding spaces that capture semantic similarities between items. Often these representations are paired, enabling experiments with analogies (pairs within the same domain) and cross-modality (pairs across domains). These experiments are based on specific assumptions about the geometry of embedding spaces, which allow finding paired items by extrapolating the positional relationships between embedding pairs in the training dataset, allowing for tasks such as finding new analogies, and multimodal zero-shot classification. In this work, we propose a metric to evaluate the similarity between paired item representations. Our proposal is built from the structural similarity between the nearest-neighbors induced graphs of each representation, and can be configured to compare spaces based on different distance metrics and on different neighborhood sizes. We demonstrate that our proposal can be used to identify similar structures at different scales, which is hard to achieve with kernel methods such as Centered Kernel Alignment (CKA). We further illustrate our method with two case studies: an analogy task using GloVe embeddings, and zero-shot classification in the CIFAR-100 dataset using CLIP embeddings. Our results show that accuracy in both analogy and zero-shot classification tasks correlates with the embedding similarity. These findings can help explain performance differences in these tasks, and may lead to improved design of paired-embedding models in the future.

arXiv:2411.07186

NatureLM-audio: an Audio-Language Foundation Model for Bioacoustics

Large language models (LLMs) prompted with text and audio represent the state of the art in various auditory tasks, including speech, music, and general audio, showing emergent abilities on unseen tasks. However, these capabilities have yet to be fully demonstrated in bioacoustics tasks, such as detecting animal vocalizations in large recordings, classifying rare and endangered species, and labeling context and behavior - tasks that are crucial for conservation, biodiversity monitoring, and the study of animal behavior. In this work, we present NatureLM-audio, the first audio-language foundation model specifically designed for bioacoustics. Our carefully curated training dataset comprises text-audio pairs spanning a diverse range of bioacoustics, speech, and music data, designed to address the challenges posed by limited annotated datasets in the field. We demonstrate successful transfer of learned representations from music and speech to bioacoustics, and our model shows promising generalization to unseen taxa and tasks. Importantly, we test NatureLM-audio on a novel benchmark (BEANS-Zero) and it sets the new state of the art (SotA) on several bioacoustics tasks, including zero-shot classification of unseen species. To advance bioacoustics research, we also open-source the code for generating training and benchmark data, as well as for training the model.

arXiv:2411.05357공식 코드 · Papers with Code

Enhancing Visual Classification using Comparative Descriptors

The performance of vision-language models (VLMs), such as CLIP, in visual classification tasks, has been enhanced by leveraging semantic knowledge from large language models (LLMs), including GPT. Recent studies have shown that in zero-shot classification tasks, descriptors incorporating additional cues, high-level concepts, or even random characters often outperform those using only the category name. In many classification tasks, while the top-1 accuracy may be relatively low, the top-5 accuracy is often significantly higher. This gap implies that most misclassifications occur among a few similar classes, highlighting the model's difficulty in distinguishing between classes with subtle differences. To address this challenge, we introduce a novel concept of comparative descriptors. These descriptors emphasize the unique features of a target class against its most similar classes, enhancing differentiation. By generating and integrating these comparative descriptors into the classification framework, we refine the semantic focus and improve classification accuracy. An additional filtering process ensures that these descriptors are closer to the image embeddings in the CLIP space, further enhancing performance. Our approach demonstrates improved accuracy and robustness in visual classification tasks by addressing the specific challenge of subtle inter-class differences.

arXiv:2411.05691

Asterisk*: Keep it Simple

This paper describes Asterisk, a compact GPT-based model for generating text embeddings. The model uses a minimalist architecture with two layers, two attention heads, and 256 embedding dimensions. By applying knowledge distillation from larger pretrained models, we explore the trade-offs between model size and performance while minimizing computational and memory requirements. The model is primarily evaluated and optimized for classification tasks, with experimental results showing its moderate performance in zero-shot classification across various downstream applications. With additional configuration, the model performance can approach or even surpass that of larger architectures on specific classification tasks.

arXiv:2411.04097공식 코드 · Papers with Code

RaVL: Discovering and Mitigating Spurious Correlations in Fine-Tuned Vision-Language Models

Fine-tuned vision-language models (VLMs) often capture spurious correlations between image features and textual attributes, resulting in degraded zero-shot performance at test time. Existing approaches for addressing spurious correlations (i) primarily operate at the global image-level rather than intervening directly on fine-grained image features and (ii) are predominantly designed for unimodal settings. In this work, we present RaVL, which takes a fine-grained perspective on VLM robustness by discovering and mitigating spurious correlations using local image features rather than operating at the global image level. Given a fine-tuned VLM, RaVL first discovers spurious correlations by leveraging a region-level clustering approach to identify precise image features contributing to zero-shot classification errors. Then, RaVL mitigates the identified spurious correlation with a novel region-aware loss function that enables the VLM to focus on relevant regions and ignore spurious relationships during fine-tuning. We evaluate RaVL on 654 VLMs with various model architectures, data domains, and learned spurious correlations. Our results show that RaVL accurately discovers (191% improvement over the closest baseline) and mitigates (8.2% improvement on worst-group image classification accuracy) spurious correlations. Qualitative evaluations on general-domain and medical-domain VLMs confirm our findings.

arXiv:2411.00246

ResiDual Transformer Alignment with Spectral Decomposition

When examined through the lens of their residual streams, a puzzling property emerges in transformer networks: residual contributions (e.g., attention heads) sometimes specialize in specific tasks or input attributes. In this paper, we analyze this phenomenon in vision transformers, focusing on the spectral geometry of residuals, and explore its implications for modality alignment in vision-language models. First, we link it to the intrinsically low-dimensional structure of visual head representations, zooming into their principal components and showing that they encode specialized roles across a wide variety of input data distributions. Then, we analyze the effect of head specialization in multimodal models, focusing on how improved alignment between text and specialized heads impacts zero-shot classification performance. This specialization-performance link consistently holds across diverse pre-training data, network sizes, and objectives, demonstrating a powerful new mechanism for boosting zero-shot classification through targeted alignment. Ultimately, we translate these insights into actionable terms by introducing ResiDual, a technique for spectral alignment of the residual stream. Much like panning for gold, it lets the noise from irrelevant unit principal components (i.e., attributes) wash away to amplify task-relevant ones. Remarkably, this dual perspective on modality alignment yields fine-tuning level performances on different data distributions while modeling an extremely interpretable and parameter-efficient transformation, as we extensively show on more than 50 (pre-trained network, dataset) pairs.

arXiv:2410.22187

Active Learning for Vision-Language Models

Pre-trained vision-language models (VLMs) like CLIP have demonstrated impressive zero-shot performance on a wide range of downstream computer vision tasks. However, there still exists a considerable performance gap between these models and a supervised deep model trained on a downstream dataset. To bridge this gap, we propose a novel active learning (AL) framework that enhances the zero-shot classification performance of VLMs by selecting only a few informative samples from the unlabeled data for annotation during training. To achieve this, our approach first calibrates the predicted entropy of VLMs and then utilizes a combination of self-uncertainty and neighbor-aware uncertainty to calculate a reliable uncertainty measure for active sample selection. Our extensive experiments show that the proposed approach outperforms existing AL approaches on several image classification datasets, and significantly enhances the zero-shot performance of VLMs.

arXiv:2410.21337

Fine-tuned Large Language Models (LLMs): Improved Prompt Injection Attacks Detection

Large language models (LLMs) are becoming a popular tool as they have significantly advanced in their capability to tackle a wide range of language-based tasks. However, LLMs applications are highly vulnerable to prompt injection attacks, which poses a critical problem. These attacks target LLMs applications through using carefully designed input prompts to divert the model from adhering to original instruction, thereby it could execute unintended actions. These manipulations pose serious security threats which potentially results in data leaks, biased outputs, or harmful responses. This project explores the security vulnerabilities in relation to prompt injection attacks. To detect whether a prompt is vulnerable or not, we follows two approaches: 1) a pre-trained LLM, and 2) a fine-tuned LLM. Then, we conduct a thorough analysis and comparison of the classification performance. Firstly, we use pre-trained XLM-RoBERTa model to detect prompt injections using test dataset without any fine-tuning and evaluate it by zero-shot classification. Then, this proposed work will apply supervised fine-tuning to this pre-trained LLM using a task-specific labeled dataset from deepset in huggingface, and this fine-tuned model achieves impressive results with 99.13\% accuracy, 100\% precision, 98.33\% recall and 99.15\% F1-score thorough rigorous experimentation and evaluation. We observe that our approach is highly efficient in detecting prompt injection attacks.

arXiv:2410.19195

Label Set Optimization via Activation Distribution Kurtosis for Zero-shot Classification with Generative Models

In-context learning (ICL) performance is known to be sensitive to the prompt design, yet the impact of class label options in zero-shot classification has been largely overlooked. This study presents the first comprehensive empirical study investigating how label option (e.g., lexical choice, order, and elaboration) influences zero-shot ICL classification performance. Our findings reveal that lexical choices for label names (e.g., agree vs.support in stance classification) play an important role, with effects also linked to label orders. An analysis of the model internal states further shows that optimal label names tend to activate fewer outlier neurons in the feed forward network. Based on this observation, we propose Label set Optimization via Activation Distribution kurtosiS (LOADS), a post-hoc approach requiring no gradient propagation. LOADS not only demonstrates effectiveness with only 100 unlabelled samples across different model types and sizes, but also shows cross-lingual transferability.

arXiv:2410.16239공식 코드 · Papers with Code

MoRE: Multi-Modal Contrastive Pre-training with Transformers on X-Rays, ECGs, and Diagnostic Report

In this paper, we introduce a novel Multi-Modal Contrastive Pre-training Framework that synergistically combines X-rays, electrocardiograms (ECGs), and radiology/cardiology reports. Our approach leverages transformers to encode these diverse modalities into a unified representation space, aiming to enhance diagnostic accuracy and facilitate comprehensive patient assessments. We utilize LoRA-Peft to significantly reduce trainable parameters in the LLM and incorporate recent linear attention dropping strategy in the Vision Transformer(ViT) for smoother attention. Furthermore, we provide novel multimodal attention explanations and retrieval for our model. To the best of our knowledge, we are the first to propose an integrated model that combines X-ray, ECG, and Radiology/Cardiology Report with this approach. By utilizing contrastive loss, MoRE effectively aligns modality-specific features into a coherent embedding, which supports various downstream tasks such as zero-shot classification and multimodal retrieval. Employing our proposed methodology, we achieve state-of-the-art (SOTA) on the Mimic-IV, CheXpert, Edema Severity, and PtbXl downstream datasets, surpassing existing multimodal approaches. Our proposed framework shows significant improvements in capturing intricate inter-modal relationships and its robustness in medical diagnosis that establishes a framework for future research in multimodal learning in the healthcare sector.

arXiv:2410.14159

Assessing Open-world Forgetting in Generative Image Model Customization

Recent advances in diffusion models have significantly enhanced image generation capabilities. However, customizing these models with new classes often leads to unintended consequences that compromise their reliability. We introduce the concept of open-world forgetting to characterize the vast scope of these unintended alterations. Our work presents the first systematic investigation into open-world forgetting in diffusion models, focusing on semantic and appearance drift of representations. Using zero-shot classification, we demonstrate that even minor model adaptations can lead to significant semantic drift affecting areas far beyond newly introduced concepts, with accuracy drops of up to 60% on previously learned concepts. Our analysis of appearance drift reveals substantial changes in texture and color distributions of generated content. To address these issues, we propose a functional regularization strategy that effectively preserves original capabilities while accommodating new concepts. Through extensive experiments across multiple datasets and evaluation metrics, we demonstrate that our approach significantly reduces both semantic and appearance drift. Our study highlights the importance of considering open-world forgetting in future research on model customization and finetuning methods.

arXiv:2410.13016공식 코드 · Papers with Code

Interpreting and Analysing CLIP's Zero-Shot Image Classification via Mutual Knowledge

Contrastive Language-Image Pretraining (CLIP) performs zero-shot image classification by mapping images and textual class representation into a shared embedding space, then retrieving the class closest to the image. This work provides a new approach for interpreting CLIP models for image classification from the lens of mutual knowledge between the two modalities. Specifically, we ask: what concepts do both vision and language CLIP encoders learn in common that influence the joint embedding space, causing points to be closer or further apart? We answer this question via an approach of textual concept-based explanations, showing their effectiveness, and perform an analysis encompassing a pool of 13 CLIP models varying in architecture, size and pretraining datasets. We explore those different aspects in relation to mutual knowledge, and analyze zero-shot predictions. Our approach demonstrates an effective and human-friendly way of understanding zero-shot classification decisions with CLIP.

arXiv:2410.13523

Can Medical Vision-Language Pre-training Succeed with Purely Synthetic Data?

Medical Vision-Language Pre-training (MedVLP) has made significant progress in enabling zero-shot tasks for medical image understanding. However, training MedVLP models typically requires large-scale datasets with paired, high-quality image-text data, which are scarce in the medical domain. Recent advancements in Large Language Models (LLMs) and diffusion models have made it possible to generate large-scale synthetic image-text pairs. This raises the question: *Can MedVLP succeed using purely synthetic data?* To address this, we use off-the-shelf generative models to create synthetic radiology reports and paired Chest X-ray (CXR) images, and propose an automated pipeline to build a diverse, high-quality synthetic dataset, enabling a rigorous study that isolates model and training settings, focusing entirely from the data perspective. Our results show that MedVLP models trained *exclusively on synthetic data* outperform those trained on real data by **3.8%** in averaged AUC on zero-shot classification. Moreover, using a combination of synthetic and real data leads to a further improvement of **9.07%**. Additionally, MedVLP models trained on synthetic or mixed data consistently outperform those trained on real data in zero-shot grounding, as well as in fine-tuned classification and segmentation tasks. Our analysis suggests MedVLP trained on well-designed synthetic data can outperform models trained on real datasets, which may be limited by low-quality samples and long-tailed distributions.

arXiv:2410.13006공식 코드 · Papers with Code

LLM Chain Ensembles for Scalable and Accurate Data Annotation

The ability of large language models (LLMs) to perform zero-shot classification makes them viable solutions for data annotation in rapidly evolving domains where quality labeled data is often scarce and costly to obtain. However, the large-scale deployment of LLMs can be prohibitively expensive. This paper introduces an LLM chain ensemble methodology that aligns multiple LLMs in a sequence, routing data subsets to subsequent models based on classification uncertainty. This approach leverages the strengths of individual LLMs within a broader system, allowing each model to handle data points where it exhibits the highest confidence, while forwarding more complex cases to potentially more robust models. Our results show that the chain ensemble method often exceeds the performance of the best individual model in the chain and achieves substantial cost savings, making LLM chain ensembles a practical and efficient solution for large-scale data annotation challenges.

arXiv:2410.11963

CtrlSynth: Controllable Image Text Synthesis for Data-Efficient Multimodal Learning

Pretraining robust vision or multimodal foundation models (e.g., CLIP) relies on large-scale datasets that may be noisy, potentially misaligned, and have long-tail distributions. Previous works have shown promising results in augmenting datasets by generating synthetic samples. However, they only support domain-specific ad hoc use cases (e.g., either image or text only, but not both), and are limited in data diversity due to a lack of fine-grained control over the synthesis process. In this paper, we design a \emph{controllable} image-text synthesis pipeline, CtrlSynth, for data-efficient and robust multimodal learning. The key idea is to decompose the visual semantics of an image into basic elements, apply user-specified control policies (e.g., remove, add, or replace operations), and recompose them to synthesize images or texts. The decompose and recompose feature in CtrlSynth allows users to control data synthesis in a fine-grained manner by defining customized control policies to manipulate the basic elements. CtrlSynth leverages the capabilities of pretrained foundation models such as large language models or diffusion models to reason and recompose basic elements such that synthetic samples are natural and composed in diverse ways. CtrlSynth is a closed-loop, training-free, and modular framework, making it easy to support different pretrained models. With extensive experiments on 31 datasets spanning different vision and vision-language tasks, we show that CtrlSynth substantially improves zero-shot classification, image-text retrieval, and compositional reasoning performance of CLIP models.

arXiv:2410.07593공식 코드 · Papers with Code

A Unified Debiasing Approach for Vision-Language Models across Modalities and Tasks

Recent advancements in Vision-Language Models (VLMs) have enabled complex multimodal tasks by processing text and image data simultaneously, significantly enhancing the field of artificial intelligence. However, these models often exhibit biases that can skew outputs towards societal stereotypes, thus necessitating debiasing strategies. Existing debiasing methods focus narrowly on specific modalities or tasks, and require extensive retraining. To address these limitations, this paper introduces Selective Feature Imputation for Debiasing (SFID), a novel methodology that integrates feature pruning and low confidence imputation (LCI) to effectively reduce biases in VLMs. SFID is versatile, maintaining the semantic integrity of outputs and costly effective by eliminating the need for retraining. Our experimental results demonstrate SFID's effectiveness across various VLMs tasks including zero-shot classification, text-to-image retrieval, image captioning, and text-to-image generation, by significantly reducing gender biases without compromising performance. This approach not only enhances the fairness of VLMs applications but also preserves their efficiency and utility across diverse scenarios.

arXiv:2410.05407

Improving Predictor Reliability with Selective Recalibration

A reliable deep learning system should be able to accurately express its confidence with respect to its predictions, a quality known as calibration. One of the most effective ways to produce reliable confidence estimates with a pre-trained model is by applying a post-hoc recalibration method. Popular recalibration methods like temperature scaling are typically fit on a small amount of data and work in the model's output space, as opposed to the more expressive feature embedding space, and thus usually have only one or a handful of parameters. However, the target distribution to which they are applied is often complex and difficult to fit well with such a function. To this end we propose \textit{selective recalibration}, where a selection model learns to reject some user-chosen proportion of the data in order to allow the recalibrator to focus on regions of the input space that can be well-captured by such a model. We provide theoretical analysis to motivate our algorithm, and test our method through comprehensive experiments on difficult medical imaging and zero-shot classification tasks. Our results show that selective recalibration consistently leads to significantly lower calibration error than a wide range of selection and recalibration baselines.

arXiv:2410.06154공식 코드 · Papers with Code

GLOV: Guided Large Language Models as Implicit Optimizers for Vision Language Models

In this work, we propose a novel method (GLOV) enabling Large Language Models (LLMs) to act as implicit Optimizers for Vision-Langugage Models (VLMs) to enhance downstream vision tasks. Our GLOV meta-prompts an LLM with the downstream task description, querying it for suitable VLM prompts (e.g., for zero-shot classification with CLIP). These prompts are ranked according to a purity measure obtained through a fitness function. In each respective optimization step, the ranked prompts are fed as in-context examples (with their accuracies) to equip the LLM with the knowledge of the type of text prompts preferred by the downstream VLM. Furthermore, we also explicitly steer the LLM generation process in each optimization step by specifically adding an offset difference vector of the embeddings from the positive and negative solutions found by the LLM, in previous optimization steps, to the intermediate layer of the network for the next generation step. This offset vector steers the LLM generation toward the type of language preferred by the downstream VLM, resulting in enhanced performance on the downstream vision tasks. We comprehensively evaluate our GLOV on 16 diverse datasets using two families of VLMs, i.e., dual-encoder (e.g., CLIP) and encoder-decoder (e.g., LLaVa) models -- showing that the discovered solutions can enhance the recognition performance by up to 15.0% and 57.5% (3.8% and 21.6% on average) for these models.

arXiv:2410.02152

An Evaluation of Large Pre-Trained Models for Gesture Recognition using Synthetic Videos

In this work, we explore the possibility of using synthetically generated data for video-based gesture recognition with large pre-trained models. We consider whether these models have sufficiently robust and expressive representation spaces to enable "training-free" classification. Specifically, we utilize various state-of-the-art video encoders to extract features for use in k-nearest neighbors classification, where the training data points are derived from synthetic videos only. We compare these results with another training-free approach -- zero-shot classification using text descriptions of each gesture. In our experiments with the RoCoG-v2 dataset, we find that using synthetic training videos yields significantly lower classification accuracy on real test videos compared to using a relatively small number of real training videos. We also observe that video backbones that were fine-tuned on classification tasks serve as superior feature extractors, and that the choice of fine-tuning data has a substantial impact on k-nearest neighbors performance. Lastly, we find that zero-shot text-based classification performs poorly on the gesture recognition task, as gestures are not easily described through natural language.