Online clothing shopping has become increasingly popular, but the high rate of returns due to size and fit issues has remained a major challenge. To address this problem, virtual try-on systems have been developed to provide customers with a more realistic and personalized way to try on clothing. In this paper, we propose a novel virtual try-on method called ClothFit, which can predict the draping shape of a garment on a target body based on the actual size of the garment and human attributes. Unlike existing try-on models, ClothFit considers the actual body proportions of the person and available cloth sizes for clothing virtualization, making it more appropriate for current online apparel outlets. The proposed method utilizes a U-Net-based network architecture that incorporates cloth and human attributes to guide the realistic virtual try-on synthesis. Specifically, we extract features from a cloth image using an auto-encoder and combine them with features from the user's height, weight, and cloth size. The features are concatenated with the features from the U-Net encoder, and the U-Net decoder synthesizes the final virtual try-on image. Our experimental results demonstrate that ClothFit can significantly improve the existing state-of-the-art methods in terms of photo-realistic virtual try-on results.
State-of-the-art (SOTA) object detection methods have succeeded in several applications at the price of relying on heavyweight neural networks, which makes them inefficient and inviable for many applications with computational resource constraints. This work presents a method to build a Convolutional Neural Network (CNN) layer by layer for object detection from user-drawn markers on discriminative regions of representative images. We address the detection of Schistosomiasis mansoni eggs in microscopy images of fecal samples, and the detection of ships in satellite images as application examples. We could create a flyweight CNN without backpropagation from very few input images. Our method explores a recent methodology, Feature Learning from Image Markers (FLIM), to build convolutional feature extractors (encoders) from marker pixels. We extend FLIM to include a single-layer adaptive decoder, whose weights vary with the input image -- a concept never explored in CNNs. Our CNN weighs thousands of times less than SOTA object detectors, being suitable for CPU execution and showing superior or equivalent performance to three methods in five measures.
Zero-shot Learning (ZSL) classification categorizes or predicts classes (labels) that are not included in the training set (unseen classes). Recent works proposed different semantic autoencoder (SAE) models where the encoder embeds a visual feature vector space into the semantic space and the decoder reconstructs the original visual feature space. The objective is to learn the embedding by leveraging a source data distribution, which can be applied effectively to a different but related target data distribution. Such embedding-based methods are prone to domain shift problems and are vulnerable to biases. We propose an integral projection-based semantic autoencoder (IP-SAE) where an encoder projects a visual feature space concatenated with the semantic space into a latent representation space. We force the decoder to reconstruct the visual-semantic data space. Due to this constraint, the visual-semantic projection function preserves the discriminatory data included inside the original visual feature space. The enriched projection forces a more precise reconstitution of the visual feature space invariant to the domain manifold. Consequently, the learned projection function is less domain-specific and alleviates the domain shift problem. Our proposed IP-SAE model consolidates a symmetric transformation function for embedding and projection, and thus, it provides transparency for interpreting generative applications in ZSL. Therefore, in addition to outperforming state-of-the-art methods considering four benchmark datasets, our analytical approach allows us to investigate distinct characteristics of generative-based methods in the unique context of zero-shot inference.
The art of mathematical reasoning stands as a fundamental pillar of intellectual progress and is a central catalyst in cultivating human ingenuity. Researchers have recently published a plethora of works centered around the task of solving Math Word Problems (MWP) $-$ a crucial stride towards general AI. These existing models are susceptible to dependency on shallow heuristics and spurious correlations to derive the solution expressions. In order to ameliorate this issue, in this paper, we propose a framework for MWP solvers based on the generation of linguistic variants of the problem text. The approach involves solving each of the variant problems and electing the predicted expression with the majority of the votes. We use DeBERTa (Decoding-enhanced BERT with disentangled attention) as the encoder to leverage its rich textual representations and enhanced mask decoder to construct the solution expressions. Furthermore, we introduce a challenging dataset, $\mathrm{P\small{ARA}\normalsize{MAWPS}}$, consisting of paraphrased, adversarial, and inverse variants of selectively sampled MWPs from the benchmark $\mathrm{M\small{AWPS}}$ dataset. We extensively experiment on this dataset along with other benchmark datasets using some baseline MWP solver models. We show that training on linguistic variants of problem statements and voting on candidate predictions improve the mathematical reasoning and robustness of the model. We make our code and data publicly available.
We introduce a framework for analyzing various types of information in an NLP Transformer. In this approach, we distinguish four layers of information: positional, syntactic, semantic, and contextual. We also argue that the common practice of adding positional information to semantic embedding is sub-optimal and propose instead a Linear-and-Add approach. Our analysis reveals an autogenetic separation of positional information through the deep layers. We show that the distilled positional components of the embedding vectors follow the path of a helix, both on the encoder side and on the decoder side. We additionally show that on the encoder side, the conceptual dimensions generate Part-of-Speech (PoS) clusters. On the decoder side, we show that a di-gram approach helps to reveal the PoS clusters of the next token. Our approach paves a way to elucidate the processing of information through the deep layers of an NLP Transformer.
As three-dimensional (3D) data acquisition devices become increasingly prevalent, the demand for 3D point cloud transmission is growing. In this study, we introduce a semantic-aware communication system for robust point cloud classification that capitalizes on the advantages of pre-trained Point-BERT models. Our proposed method comprises four main components: the semantic encoder, channel encoder, channel decoder, and semantic decoder. By employing a two-stage training strategy, our system facilitates efficient and adaptable learning tailored to the specific classification tasks. The results show that the proposed system achieves classification accuracy of over 89\% when SNR is higher than 10 dB and still maintains accuracy above 66.6\% even at SNR of 4 dB. Compared to the existing method, our approach performs at 0.8\% to 48\% better across different SNR values, demonstrating robustness to channel noise. Our system also achieves a balance between accuracy and speed, being computationally efficient while maintaining high classification performance under noisy channel conditions. This adaptable and resilient approach holds considerable promise for a wide array of 3D scene understanding applications, effectively addressing the challenges posed by channel noise.
End-to-end design of communication systems using deep autoencoders (AEs) is gaining attention due to its flexibility and excellent performance. Besides single-user transmission, AE-based design is recently explored in multi-user setup, e.g., for designing constellations for non-orthogonal multiple access (NOMA). In this paper, we further advance the design of AE-based downlink NOMA by introducing weighted loss function in the AE training. By changing the weight coefficients, one can flexibly tune the constellation design to balance error probability of different users, without relying on explicit information about their channel quality. Combined with the SICNet decoder, we demonstrate a significant improvement in achievable levels and flexible control of error probability of different users using the proposed weighted AE-based framework.
Knowledge graphs (KGs), containing many entity-relation-entity triples, provide rich information for downstream applications. Although extracting triples from unstructured texts has been widely explored, most of them require a large number of labeled instances. The performance will drop dramatically when only few labeled data are available. To tackle this problem, we propose the Mutually Guided Few-shot learning framework for Relational Triple Extraction (MG-FTE). Specifically, our method consists of an entity-guided relation proto-decoder to classify the relations firstly and a relation-guided entity proto-decoder to extract entities based on the classified relations. To draw the connection between entity and relation, we design a proto-level fusion module to boost the performance of both entity extraction and relation classification. Moreover, a new cross-domain few-shot triple extraction task is introduced. Extensive experiments show that our method outperforms many state-of-the-art methods by 12.6 F1 score on FewRel 1.0 (single-domain) and 20.5 F1 score on FewRel 2.0 (cross-domain).
Recently, foundation models have been introduced demonstrating various tasks in the field of computer vision. These models such as Segment Anything Model (SAM) are generalized models trained using huge datasets. Currently, ongoing research focuses on exploring the effective utilization of these generalized models for specific domains, such as medical imaging. However, in medical imaging, the lack of training samples due to privacy concerns and other factors presents a major challenge for applying these generalized models to medical image segmentation task. To address this issue, the effective fine tuning of these models is crucial to ensure their optimal utilization. In this study, we propose to combine a complementary Convolutional Neural Network (CNN) along with the standard SAM network for medical image segmentation. To reduce the burden of fine tuning large foundation model and implement cost-efficient trainnig scheme, we focus only on fine-tuning the additional CNN network and SAM decoder part. This strategy significantly reduces trainnig time and achieves competitive results on publicly available dataset. The code is available at https://github.com/11yxk/SAM-LST.
This work proposes a novel face-swapping framework FlowFace++, utilizing explicit semantic flow supervision and end-to-end architecture to facilitate shape-aware face-swapping. Specifically, our work pretrains a facial shape discriminator to supervise the face swapping network. The discriminator is shape-aware and relies on a semantic flow-guided operation to explicitly calculate the shape discrepancies between the target and source faces, thus optimizing the face swapping network to generate highly realistic results. The face swapping network is a stack of a pre-trained face-masked autoencoder (MAE), a cross-attention fusion module, and a convolutional decoder. The MAE provides a fine-grained facial image representation space, which is unified for the target and source faces and thus facilitates final realistic results. The cross-attention fusion module carries out the source-to-target face swapping in a fine-grained latent space while preserving other attributes of the target image (e.g. expression, head pose, hair, background, illumination, etc). Lastly, the convolutional decoder further synthesizes the swapping results according to the face-swapping latent embedding from the cross-attention fusion module. Extensive quantitative and qualitative experiments on in-the-wild faces demonstrate that our FlowFace++ outperforms the state-of-the-art significantly, particularly while the source face is obstructed by uneven lighting or angle offset.
The emergence of different sensors (Near-Infrared, Depth, etc.) is a remedy for the limited application scenarios of traditional RGB camera. The RGB-X tasks, which rely on RGB input and another type of data input to resolve specific problems, have become a popular research topic in multimedia. A crucial part in two-branch RGB-X deep neural networks is how to fuse information across modalities. Given the tremendous information inside RGB-X networks, previous works typically apply naive fusion (e.g., average or max fusion) or only focus on the feature fusion at the same scale(s). While in this paper, we propose a novel method called RXFOOD for the fusion of features across different scales within the same modality branch and from different modality branches simultaneously in a unified attention mechanism. An Energy Exchange Module is designed for the interaction of each feature map's energy matrix, who reflects the inter-relationship of different positions and different channels inside a feature map. The RXFOOD method can be easily incorporated to any dual-branch encoder-decoder network as a plug-in module, and help the original backbone network better focus on important positions and channels for object of interest detection. Experimental results on RGB-NIR salient object detection, RGB-D salient object detection, and RGBFrequency image manipulation detection demonstrate the clear effectiveness of the proposed RXFOOD.
Change detection is of fundamental importance when analyzing data streams. Detecting changes both quickly and accurately enables monitoring and prediction systems to react, e.g., by issuing an alarm or by updating a learning algorithm. However, detecting changes is challenging when observations are high-dimensional. In high-dimensional data, change detectors should not only be able to identify when changes happen, but also in which subspace they occur. Ideally, one should also quantify how severe they are. Our approach, ABCD, has these properties. ABCD learns an encoder-decoder model and monitors its accuracy over a window of adaptive size. ABCD derives a change score based on Bernstein's inequality to detect deviations in terms of accuracy, which indicate changes. Our experiments demonstrate that ABCD outperforms its best competitor by up to 20% in F1-score on average. It can also accurately estimate changes' subspace, together with a severity measure that correlates with the ground truth.
In this paper, we study representation learning in partially observable Markov Decision Processes (POMDPs), where the agent learns a decoder function that maps a series of high-dimensional raw observations to a compact representation and uses it for more efficient exploration and planning. We focus our attention on the sub-classes of \textit{$\gamma$-observable} and \textit{decodable POMDPs}, for which it has been shown that statistically tractable learning is possible, but there has not been any computationally efficient algorithm. We first present an algorithm for decodable POMDPs that combines maximum likelihood estimation (MLE) and optimism in the face of uncertainty (OFU) to perform representation learning and achieve efficient sample complexity, while only calling supervised learning computational oracles. We then show how to adapt this algorithm to also work in the broader class of $\gamma$-observable POMDPs.
In this work, we demonstrate the application of a first-order Taylor expansion to approximate a generic function $F: R^{n \times m} \to R^{n \times m}$ and utilize it in language modeling. To enhance the basic Taylor expansion, we introduce iteration and piecewise modeling, leading us to name the algorithm the Iterative Piecewise Affine (IPA) approximation. The final algorithm exhibits interesting resemblances to the Transformers decoder architecture. By comparing parameter arrangements in IPA and Transformers, we observe a strikingly similar performance, with IPA outperforming Transformers by 1.5\% in the next token prediction task with cross-entropy loss for smaller sequence lengths.
We propose an efficient abnormal event detection model based on a lightweight masked auto-encoder (AE) applied at the video frame level. The novelty of the proposed model is threefold. First, we introduce an approach to weight tokens based on motion gradients, thus shifting the focus from the static background scene to the foreground objects. Second, we integrate a teacher decoder and a student decoder into our architecture, leveraging the discrepancy between the outputs given by the two decoders to improve anomaly detection. Third, we generate synthetic abnormal events to augment the training videos, and task the masked AE model to jointly reconstruct the original frames (without anomalies) and the corresponding pixel-level anomaly maps. Our design leads to an efficient and effective model, as demonstrated by the extensive experiments carried out on four benchmarks: Avenue, ShanghaiTech, UBnormal and UCSD Ped2. The empirical results show that our model achieves an excellent trade-off between speed and accuracy, obtaining competitive AUC scores, while processing 1655 FPS. Hence, our model is between 8 and 70 times faster than competing methods. We also conduct an ablation study to justify our design. Our code is freely available at: https://github.com/ristea/aed-mae.
Scene text removal (STR) aims at replacing text strokes in natural scenes with visually coherent backgrounds. Recent STR approaches rely on iterative refinements or explicit text masks, resulting in high complexity and sensitivity to the accuracy of text localization. Moreover, most existing STR methods adopt convolutional architectures while the potential of vision Transformers (ViTs) remains largely unexplored. In this paper, we propose a simple-yet-effective ViT-based text eraser, dubbed ViTEraser. Following a concise encoder-decoder framework, ViTEraser can easily incorporate various ViTs to enhance long-range modeling. Specifically, the encoder hierarchically maps the input image into the hidden space through ViT blocks and patch embedding layers, while the decoder gradually upsamples the hidden features to the text-erased image with ViT blocks and patch splitting layers. As ViTEraser implicitly integrates text localization and inpainting, we propose a novel end-to-end pretraining method, termed SegMIM, which focuses the encoder and decoder on the text box segmentation and masked image modeling tasks, respectively. Experimental results demonstrate that ViTEraser with SegMIM achieves state-of-the-art performance on STR by a substantial margin and exhibits strong generalization ability when extended to other tasks, \textit{e.g.}, tampered scene text detection. Furthermore, we comprehensively explore the architecture, pretraining, and scalability of the ViT-based encoder-decoder for STR, which provides deep insights into the application of ViT to the STR field. Code is available at https://github.com/shannanyinxiang/ViTEraser.
In recent years, the integration of Large Language Models (LLMs) into recommender systems has garnered interest among both practitioners and researchers. Despite this interest, the field is still emerging, and the lack of open-source R&D platforms may impede the exploration of LLM-based recommendations. This paper introduces OpenP5, an open-source platform designed as a resource to facilitate the development, training, and evaluation of LLM-based generative recommender systems for research purposes. The platform is implemented using encoder-decoder LLMs (e.g., T5) and decoder-only LLMs (e.g., Llama-2) across 10 widely recognized public datasets, catering to two fundamental recommendation tasks: sequential and straightforward recommendations. Recognizing the crucial role of item IDs in LLM-based recommendations, we have also incorporated three item indexing methods within the OpenP5 platform: random indexing, sequential indexing and collaborative indexing. Built on the Transformers library, the platform facilitates easy customization of LLM-based recommendations for users. OpenP5 boasts a range of features including extensible data processing, task-centric optimization, comprehensive datasets and checkpoints, efficient acceleration, and standardized evaluations, making it a valuable tool for the implementation and evaluation of LLM-based recommender systems. The open-source code and pre-trained checkpoints for the OpenP5 library are publicly available at https://github.com/agiresearch/OpenP5.
We present a novel typical-to-atypical voice conversion approach (DuTa-VC), which (i) can be trained with nonparallel data (ii) first introduces diffusion probabilistic model (iii) preserves the target speaker identity (iv) is aware of the phoneme duration of the target speaker. DuTa-VC consists of three parts: an encoder transforms the source mel-spectrogram into a duration-modified speaker-independent mel-spectrogram, a decoder performs the reverse diffusion to generate the target mel-spectrogram, and a vocoder is applied to reconstruct the waveform. Objective evaluations conducted on the UASpeech show that DuTa-VC is able to capture severity characteristics of dysarthric speech, reserves speaker identity, and significantly improves dysarthric speech recognition as a data augmentation. Subjective evaluations by two expert speech pathologists validate that DuTa-VC can preserve the severity and type of dysarthria of the target speakers in the synthesized speech.
The Streaming Unmixing and Recognition Transducer (SURT) model was proposed recently as an end-to-end approach for continuous, streaming, multi-talker speech recognition (ASR). Despite impressive results on multi-turn meetings, SURT has notable limitations: (i) it suffers from leakage and omission related errors; (ii) it is computationally expensive, due to which it has not seen adoption in academia; and (iii) it has only been evaluated on synthetic mixtures. In this work, we propose several modifications to the original SURT which are carefully designed to fix the above limitations. In particular, we (i) change the unmixing module to a mask estimator that uses dual-path modeling, (ii) use a streaming zipformer encoder and a stateless decoder for the transducer, (iii) perform mixture simulation using force-aligned subsegments, (iv) pre-train the transducer on single-speaker data, (v) use auxiliary objectives in the form of masking loss and encoder CTC loss, and (vi) perform domain adaptation for far-field recognition. We show that our modifications allow SURT 2.0 to outperform its predecessor in terms of multi-talker ASR results, while being efficient enough to train with academic resources. We conduct our evaluations on 3 publicly available meeting benchmarks -- LibriCSS, AMI, and ICSI, where our best model achieves WERs of 16.9%, 44.6% and 32.2%, respectively, on far-field unsegmented recordings. We release training recipes and pre-trained models: https://sites.google.com/view/surt2.
Recently, speech-to-text translation has attracted more and more attention and many studies have emerged rapidly. In this paper, we present a comprehensive survey on direct speech translation aiming to summarize the current state-of-the-art techniques. First, we categorize the existing research work into three directions based on the main challenges -- modeling burden, data scarcity, and application issues. To tackle the problem of modeling burden, two main structures have been proposed, encoder-decoder framework (Transformer and the variants) and multitask frameworks. For the challenge of data scarcity, recent work resorts to many sophisticated techniques, such as data augmentation, pre-training, knowledge distillation, and multilingual modeling. We analyze and summarize the application issues, which include real-time, segmentation, named entity, gender bias, and code-switching. Finally, we discuss some promising directions for future work.