Revisiting MLLMs: An In-Depth Analysis of Image Classification Abilities

With the rapid advancement of Multimodal Large Language Models (MLLMs), a variety of benchmarks have been introduced to evaluate their capabilities. While most evaluations have focused on complex tasks such as scientific comprehension and visual reasoning, little attention has been given to assessing their fundamental image classification abilities. In this paper, we address this gap by thoroughly revisiting the MLLMs with an in-depth analysis of image classification. Specifically, building on established datasets, we examine a broad spectrum of scenarios, from general classification tasks (e.g., ImageNet, ObjectNet) to more fine-grained categories such as bird and food classification. Our findings reveal that the most recent MLLMs can match or even outperform CLIP-style vision-language models on several datasets, challenging the previous assumption that MLLMs are bad at image classification \cite{VLMClassifier}. To understand the factors driving this improvement, we conduct an in-depth analysis of the network architecture, data selection, and training recipe used in public MLLMs. Our results attribute this success to advancements in language models and the diversity of training data sources. Based on these observations, we further analyze and attribute the potential reasons to conceptual knowledge transfer and enhanced exposure of target concepts, respectively. We hope our findings will offer valuable insights for future research on MLLMs and their evaluation in image classification tasks.

Descriptive Caption Enhancement with Visual Specialists for Multimodal Perception

Training Large Multimodality Models (LMMs) relies on descriptive image caption that connects image and language. Existing methods either distill the caption from the LMM models or construct the captions from the internet images or by human. We propose to leverage off-the-shelf visual specialists, which were trained from annotated images initially not for image captioning, for enhancing the image caption. Our approach, named DCE, explores object low-level and fine-grained attributes (e.g., depth, emotion and fine-grained categories) and object relations (e.g., relative location and human-object-interaction (HOI)), and combine the attributes into the descriptive caption. Experiments demonstrate that such visual specialists are able to improve the performance for visual understanding tasks as well as reasoning that benefits from more accurate visual understanding. We will release the source code and the pipeline so that other visual specialists are easily combined into the pipeline. The complete source code of DCE pipeline and datasets will be available at \url{https://github.com/syp2ysy/DCE}.

Trustworthy Multimodal Fusion for Sentiment Analysis in Ordinal Sentiment Space

Multimodal video sentiment analysis aims to integrate multiple modal information to analyze the opinions and attitudes of speakers. Most previous work focuses on exploring the semantic interactions of intra- and inter-modality. However, these works ignore the reliability of multimodality, i.e., modalities tend to contain noise, semantic ambiguity, missing modalities, etc. In addition, previous multimodal approaches treat different modalities equally, largely ignoring their different contributions. Furthermore, existing multimodal sentiment analysis methods directly regress sentiment scores without considering ordinal relationships within sentiment categories, with limited performance. To address the aforementioned problems, we propose a trustworthy multimodal sentiment ordinal network (TMSON) to improve performance in sentiment analysis. Specifically, we first devise a unimodal feature extractor for each modality to obtain modality-specific features. Then, an uncertainty distribution estimation network is customized, which estimates the unimodal uncertainty distributions. Next, Bayesian fusion is performed on the learned unimodal distributions to obtain multimodal distributions for sentiment prediction. Finally, an ordinal-aware sentiment space is constructed, where ordinal regression is used to constrain the multimodal distributions. Our proposed TMSON outperforms baselines on multimodal sentiment analysis tasks, and empirical results demonstrate that TMSON is capable of reducing uncertainty to obtain more robust predictions.

PromptAD: Learning Prompts with only Normal Samples for Few-Shot Anomaly Detection

The vision-language model has brought great improvement to few-shot industrial anomaly detection, which usually needs to design of hundreds of prompts through prompt engineering. For automated scenarios, we first use conventional prompt learning with many-class paradigm as the baseline to automatically learn prompts but found that it can not work well in one-class anomaly detection. To address the above problem, this paper proposes a one-class prompt learning method for few-shot anomaly detection, termed PromptAD. First, we propose semantic concatenation which can transpose normal prompts into anomaly prompts by concatenating normal prompts with anomaly suffixes, thus constructing a large number of negative samples used to guide prompt learning in one-class setting. Furthermore, to mitigate the training challenge caused by the absence of anomaly images, we introduce the concept of explicit anomaly margin, which is used to explicitly control the margin between normal prompt features and anomaly prompt features through a hyper-parameter. For image-level/pixel-level anomaly detection, PromptAD achieves first place in 11/12 few-shot settings on MVTec and VisA.

X2-Softmax: Margin Adaptive Loss Function for Face Recognition

Learning the discriminative features of different faces is an important task in face recognition. By extracting face features in neural networks, it becomes easy to measure the similarity of different face images, which makes face recognition possible. To enhance the neural network's face feature separability, incorporating an angular margin during training is common practice. State-of-the-art loss functions CosFace and ArcFace apply fixed margins between weights of classes to enhance the inter-class separation of face features. Since the distribution of samples in the training set is imbalanced, similarities between different identities are unequal. Therefore, using an inappropriately fixed angular margin may lead to the problem that the model is difficult to converge or the face features are not discriminative enough. It is more in line with our intuition that the margins are angular adaptive, which could increase with the angles between classes growing. In this paper, we propose a new angular margin loss named X2-Softmax. X2-Softmax loss has adaptive angular margins, which provide the margin that increases with the angle between different classes growing. The angular adaptive margin ensures model flexibility and effectively improves the effect of face recognition. We have trained the neural network with X2-Softmax loss on the MS1Mv3 dataset and tested it on several evaluation benchmarks to demonstrate the effectiveness and superiority of our loss function.

HARQ-IR Aided Short Packet Communications: BLER Analysis and Throughput Maximization

This paper introduces hybrid automatic repeat request with incremental redundancy (HARQ-IR) to boost the reliability of short packet communications. The finite blocklength information theory and correlated decoding events tremendously preclude the analysis of average block error rate (BLER). Fortunately, the recursive form of average BLER motivates us to calculate its value through the trapezoidal approximation and Gauss-Laguerre quadrature. Moreover, the asymptotic analysis is performed to derive a simple expression for the average BLER at high signal-to-noise ratio (SNR). Then, we study the maximization of long term average throughput (LTAT) via power allocation meanwhile ensuring the power and the BLER constraints. For tractability, the asymptotic BLER is employed to solve the problem through geometric programming (GP). However, the GP-based solution underestimates the LTAT at low SNR due to a large approximation error in this case. Alternatively, we also develop a deep reinforcement learning (DRL)-based framework to learn power allocation policy. In particular, the optimization problem is transformed into a constrained Markov decision process, which is solved by integrating deep deterministic policy gradient (DDPG) with subgradient method. The numerical results finally demonstrate that the DRL-based method outperforms the GP-based one at low SNR, albeit at the cost of increasing computational burden.

DrivingDiffusion: Layout-Guided multi-view driving scene video generation with latent diffusion model

With the increasing popularity of autonomous driving based on the powerful and unified bird's-eye-view (BEV) representation, a demand for high-quality and large-scale multi-view video data with accurate annotation is urgently required. However, such large-scale multi-view data is hard to obtain due to expensive collection and annotation costs. To alleviate the problem, we propose a spatial-temporal consistent diffusion framework DrivingDiffusion, to generate realistic multi-view videos controlled by 3D layout. There are three challenges when synthesizing multi-view videos given a 3D layout: How to keep 1) cross-view consistency and 2) cross-frame consistency? 3) How to guarantee the quality of the generated instances? Our DrivingDiffusion solves the problem by cascading the multi-view single-frame image generation step, the single-view video generation step shared by multiple cameras, and post-processing that can handle long video generation. In the multi-view model, the consistency of multi-view images is ensured by information exchange between adjacent cameras. In the temporal model, we mainly query the information that needs attention in subsequent frame generation from the multi-view images of the first frame. We also introduce the local prompt to effectively improve the quality of generated instances. In post-processing, we further enhance the cross-view consistency of subsequent frames and extend the video length by employing temporal sliding window algorithm. Without any extra cost, our model can generate large-scale realistic multi-camera driving videos in complex urban scenes, fueling the downstream driving tasks. The code will be made publicly available.

USL-Net: Uncertainty Self-Learning Network for Unsupervised Skin Lesion Segmentation

Unsupervised skin lesion segmentation offers several benefits, including conserving expert human resources, reducing discrepancies due to subjective human labeling, and adapting to novel environments. However, segmenting dermoscopic images without manual labeling guidance presents significant challenges due to dermoscopic image artifacts such as hair noise, blister noise, and subtle edge differences. To address these challenges, we introduce an innovative Uncertainty Self-Learning Network (USL-Net) designed for skin lesion segmentation. The USL-Net can effectively segment a range of lesions, eliminating the need for manual labeling guidance. Initially, features are extracted using contrastive learning, followed by the generation of Class Activation Maps (CAMs) as saliency maps using these features. The different CAM locations correspond to the importance of the lesion region based on their saliency. High-saliency regions in the map serve as pseudo-labels for lesion regions while low-saliency regions represent the background. However, intermediate regions can be hard to classify, often due to their proximity to lesion edges or interference from hair or blisters. Rather than risk potential pseudo-labeling errors or learning confusion by forcefully classifying these regions, we consider them as uncertainty regions, exempting them from pseudo-labeling and allowing the network to self-learn. Further, we employ connectivity detection and centrality detection to refine foreground pseudo-labels and reduce noise-induced errors. The application of cycle refining enhances performance further. Our method underwent thorough experimental validation on the ISIC-2017, ISIC-2018, and PH2 datasets, demonstrating that its performance is on par with weakly supervised and supervised methods, and exceeds that of other existing unsupervised methods.

Over-the-Air Aggregation for Federated Learning: Waveform Superposition and Prototype Validation

In this paper, we develop an orthogonal-frequency-division-multiplexing (OFDM)-based over-the-air (OTA) aggregation solution for wireless federated learning (FL). In particular, the local gradients in massive IoT devices are modulated by an analog waveform and are then transmitted using the same wireless resources. To this end, achieving perfect waveform superposition is the key challenge, which is difficult due to the existence of frame timing offset (TO) and carrier frequency offset (CFO). In order to address these issues, we propose a two-stage waveform pre-equalization technique with a customized multiple access protocol that can estimate and then mitigate the TO and CFO for the OTA aggregation. Based on the proposed solution, we develop a hardware transceiver and application software to train a real-world FL task, which learns a deep neural network to predict the received signal strength with global positioning system information. Experiments verify that the proposed OTA aggregation solution can achieve comparable performance to offline learning procedures with high prediction accuracy.

State estimation under non-Gaussian Levy noise: A modified Kalman filtering method

The Kalman filter is extensively used for state estimation for linear systems under Gaussian noise. When non-Gaussian L\'evy noise is present, the conventional Kalman filter may fail to be effective due to the fact that the non-Gaussian L\'evy noise may have infinite variance. A modified Kalman filter for linear systems with non-Gaussian L\'evy noise is devised. It works effectively with reasonable computational cost. Simulation results are presented to illustrate this non-Gaussian filtering method.