Continuous Speculative Decoding for Autoregressive Image Generation

Continuous-valued Autoregressive (AR) image generation models have demonstrated notable superiority over their discrete-token counterparts, showcasing considerable reconstruction quality and higher generation fidelity. However, the computational demands of the autoregressive framework result in significant inference overhead. While speculative decoding has proven effective in accelerating Large Language Models (LLMs), their adaptation to continuous-valued visual autoregressive models remains unexplored. This work generalizes the speculative decoding algorithm from discrete tokens to continuous space. By analyzing the intrinsic properties of output distribution, we establish a tailored acceptance criterion for the diffusion distributions prevalent in such models. To overcome the inconsistency that occurred in speculative decoding output distributions, we introduce denoising trajectory alignment and token pre-filling methods. Additionally, we identify the hard-to-sample distribution in the rejection phase. To mitigate this issue, we propose a meticulous acceptance-rejection sampling method with a proper upper bound, thereby circumventing complex integration. Experimental results show that our continuous speculative decoding achieves a remarkable $2.33\times$ speed-up on off-the-shelf models while maintaining the output distribution. Codes will be available at https://github.com/MarkXCloud/CSpD

A Survey of Low-shot Vision-Language Model Adaptation via Representer Theorem

The advent of pre-trained vision-language foundation models has revolutionized the field of zero/few-shot (i.e., low-shot) image recognition. The key challenge to address under the condition of limited training data is how to fine-tune pre-trained vision-language models in a parameter-efficient manner. Previously, numerous approaches tackling this challenge have been proposed. Meantime, a few survey papers are also published to summarize these works. However, there still lacks a unified computational framework to integrate existing methods together, identify their nature and support in-depth comparison. As such, this survey paper first proposes a unified computational framework from the perspective of Representer Theorem and then derives many of the existing methods by specializing this framework. Thereafter, a comparative analysis is conducted to uncover the differences and relationships between existing methods. Based on the analyses, some possible variants to improve the existing works are presented. As a demonstration, we extend existing methods by modeling inter-class correlation between representers in reproducing kernel Hilbert space (RKHS), which is implemented by exploiting the closed-form solution of kernel ridge regression. Extensive experiments on 11 datasets are conducted to validate the effectiveness of this method. Toward the end of this paper, we discuss the limitations and provide further research directions.

Calibrated Cache Model for Few-Shot Vision-Language Model Adaptation

Cache-based approaches stand out as both effective and efficient for adapting vision-language models (VLMs). Nonetheless, the existing cache model overlooks three crucial aspects. 1) Pre-trained VLMs are mainly optimized for image-text similarity, neglecting the importance of image-image similarity, leading to a gap between pre-training and adaptation. 2) The current cache model is based on the Nadaraya-Watson (N-W) estimator, which disregards the intricate relationships among training samples while constructing weight function. 3) Under the condition of limited samples, the logits generated by cache model are of high uncertainty, directly using these logits without accounting for the confidence could be problematic. This work presents three calibration modules aimed at addressing the above challenges. Similarity Calibration refines the image-image similarity by using unlabeled images. We add a learnable projection layer with residual connection on top of the pre-trained image encoder of CLIP and optimize the parameters by minimizing self-supervised contrastive loss. Weight Calibration introduces a precision matrix into the weight function to adequately model the relation between training samples, transforming the existing cache model to a Gaussian Process (GP) regressor, which could be more accurate than N-W estimator. Confidence Calibration leverages the predictive variances computed by GP Regression to dynamically re-scale the logits of cache model, ensuring that the cache model's outputs are appropriately adjusted based on their confidence levels. Besides, to reduce the high complexity of GPs, we further propose a group-based learning strategy. Integrating the above designs, we propose both training-free and training-required variants. Extensive experiments on 11 few-shot classification datasets validate that the proposed methods can achieve state-of-the-art performance.

Draw an Audio: Leveraging Multi-Instruction for Video-to-Audio Synthesis

Foley is a term commonly used in filmmaking, referring to the addition of daily sound effects to silent films or videos to enhance the auditory experience. Video-to-Audio (V2A), as a particular type of automatic foley task, presents inherent challenges related to audio-visual synchronization. These challenges encompass maintaining the content consistency between the input video and the generated audio, as well as the alignment of temporal and loudness properties within the video. To address these issues, we construct a controllable video-to-audio synthesis model, termed Draw an Audio, which supports multiple input instructions through drawn masks and loudness signals. To ensure content consistency between the synthesized audio and target video, we introduce the Mask-Attention Module (MAM), which employs masked video instruction to enable the model to focus on regions of interest. Additionally, we implement the Time-Loudness Module (TLM), which uses an auxiliary loudness signal to ensure the synthesis of sound that aligns with the video in both loudness and temporal dimensions. Furthermore, we have extended a large-scale V2A dataset, named VGGSound-Caption, by annotating caption prompts. Extensive experiments on challenging benchmarks across two large-scale V2A datasets verify Draw an Audio achieves the state-of-the-art. Project page: https://yannqi.github.io/Draw-an-Audio/.

AVESFormer: Efficient Transformer Design for Real-Time Audio-Visual Segmentation

Recently, transformer-based models have demonstrated remarkable performance on audio-visual segmentation (AVS) tasks. However, their expensive computational cost makes real-time inference impractical. By characterizing attention maps of the network, we identify two key obstacles in AVS models: 1) attention dissipation, corresponding to the over-concentrated attention weights by Softmax within restricted frames, and 2) inefficient, burdensome transformer decoder, caused by narrow focus patterns in early stages. In this paper, we introduce AVESFormer, the first real-time Audio-Visual Efficient Segmentation transformer that achieves fast, efficient and light-weight simultaneously. Our model leverages an efficient prompt query generator to correct the behaviour of cross-attention. Additionally, we propose ELF decoder to bring greater efficiency by facilitating convolutions suitable for local features to reduce computational burdens. Extensive experiments demonstrate that our AVESFormer significantly enhances model performance, achieving 79.9% on S4, 57.9% on MS3 and 31.2% on AVSS, outperforming previous state-of-the-art and achieving an excellent trade-off between performance and speed. Code can be found at https://github.com/MarkXCloud/AVESFormer.git.

AddressCLIP: Empowering Vision-Language Models for City-wide Image Address Localization

In this study, we introduce a new problem raised by social media and photojournalism, named Image Address Localization (IAL), which aims to predict the readable textual address where an image was taken. Existing two-stage approaches involve predicting geographical coordinates and converting them into human-readable addresses, which can lead to ambiguity and be resource-intensive. In contrast, we propose an end-to-end framework named AddressCLIP to solve the problem with more semantics, consisting of two key ingredients: i) image-text alignment to align images with addresses and scene captions by contrastive learning, and ii) image-geography matching to constrain image features with the spatial distance in terms of manifold learning. Additionally, we have built three datasets from Pittsburgh and San Francisco on different scales specifically for the IAL problem. Experiments demonstrate that our approach achieves compelling performance on the proposed datasets and outperforms representative transfer learning methods for vision-language models. Furthermore, extensive ablations and visualizations exhibit the effectiveness of the proposed method. The datasets and source code are available at https://github.com/xsx1001/AddressCLIP.

SegICL: A Universal In-context Learning Framework for Enhanced Segmentation in Medical Imaging

Medical image segmentation models adapting to new tasks in a training-free manner through in-context learning is an exciting advancement. Universal segmentation models aim to generalize across the diverse modality of medical images, yet their effectiveness often diminishes when applied to out-of-distribution (OOD) data modalities and tasks, requiring intricate fine-tuning of model for optimal performance. For addressing this challenge, we introduce SegICL, a novel approach leveraging In-Context Learning (ICL) for image segmentation. Unlike existing methods, SegICL has the capability to employ text-guided segmentation and conduct in-context learning with a small set of image-mask pairs, eliminating the need for training the model from scratch or fine-tuning for OOD tasks (including OOD modality and dataset). Extensive experimental validation of SegICL demonstrates a positive correlation between the number of prompt samples and segmentation performance on OOD modalities and tasks. This indicates that SegICL effectively address new segmentation tasks based on contextual information. Additionally, SegICL also exhibits comparable segmentation performance to mainstream models on OOD and in-distribution tasks. Our code will be released soon.

Weak Distribution Detectors Lead to Stronger Generalizability of Vision-Language Prompt Tuning

We propose a generalized method for boosting the generalization ability of pre-trained vision-language models (VLMs) while fine-tuning on downstream few-shot tasks. The idea is realized by exploiting out-of-distribution (OOD) detection to predict whether a sample belongs to a base distribution or a novel distribution and then using the score generated by a dedicated competition based scoring function to fuse the zero-shot and few-shot classifier. The fused classifier is dynamic, which will bias towards the zero-shot classifier if a sample is more likely from the distribution pre-trained on, leading to improved base-to-novel generalization ability. Our method is performed only in test stage, which is applicable to boost existing methods without time-consuming re-training. Extensive experiments show that even weak distribution detectors can still improve VLMs' generalization ability. Specifically, with the help of OOD detectors, the harmonic mean of CoOp and ProGrad increase by 2.6 and 1.5 percentage points over 11 recognition datasets in the base-to-novel setting.

Reusable Architecture Growth for Continual Stereo Matching

The remarkable performance of recent stereo depth estimation models benefits from the successful use of convolutional neural networks to regress dense disparity. Akin to most tasks, this needs gathering training data that covers a number of heterogeneous scenes at deployment time. However, training samples are typically acquired continuously in practical applications, making the capability to learn new scenes continually even more crucial. For this purpose, we propose to perform continual stereo matching where a model is tasked to 1) continually learn new scenes, 2) overcome forgetting previously learned scenes, and 3) continuously predict disparities at inference. We achieve this goal by introducing a Reusable Architecture Growth (RAG) framework. RAG leverages task-specific neural unit search and architecture growth to learn new scenes continually in both supervised and self-supervised manners. It can maintain high reusability during growth by reusing previous units while obtaining good performance. Additionally, we present a Scene Router module to adaptively select the scene-specific architecture path at inference. Comprehensive experiments on numerous datasets show that our framework performs impressively in various weather, road, and city circumstances and surpasses the state-of-the-art methods in more challenging cross-dataset settings. Further experiments also demonstrate the adaptability of our method to unseen scenes, which can facilitate end-to-end stereo architecture learning and practical deployment.

Enhancing Visual Continual Learning with Language-Guided Supervision

Continual learning (CL) aims to empower models to learn new tasks without forgetting previously acquired knowledge. Most prior works concentrate on the techniques of architectures, replay data, regularization, \etc. However, the category name of each class is largely neglected. Existing methods commonly utilize the one-hot labels and randomly initialize the classifier head. We argue that the scarce semantic information conveyed by the one-hot labels hampers the effective knowledge transfer across tasks. In this paper, we revisit the role of the classifier head within the CL paradigm and replace the classifier with semantic knowledge from pretrained language models (PLMs). Specifically, we use PLMs to generate semantic targets for each class, which are frozen and serve as supervision signals during training. Such targets fully consider the semantic correlation between all classes across tasks. Empirical studies show that our approach mitigates forgetting by alleviating representation drifting and facilitating knowledge transfer across tasks. The proposed method is simple to implement and can seamlessly be plugged into existing methods with negligible adjustments. Extensive experiments based on eleven mainstream baselines demonstrate the effectiveness and generalizability of our approach to various protocols. For example, under the class-incremental learning setting on ImageNet-100, our method significantly improves the Top-1 accuracy by 3.2\% to 6.1\% while reducing the forgetting rate by 2.6\% to 13.1\%.