Moir\'e patterns are commonly seen when taking photos of screens. Camera devices usually have limited hardware performance but take high-resolution photos. However, users are sensitive to the photo processing time, which presents a hardly considered challenge of efficiency for demoir\'eing methods. To balance the network speed and quality of results, we propose a \textbf{F}ully \textbf{C}onnected en\textbf{C}oder-de\textbf{C}oder based \textbf{D}emoir\'eing \textbf{Net}work (FC3DNet). FC3DNet utilizes features with multiple scales in each stage of the decoder for comprehensive information, which contains long-range patterns as well as various local moir\'e styles that both are crucial aspects in demoir\'eing. Besides, to make full use of multiple features, we design a Multi-Feature Multi-Attention Fusion (MFMAF) module to weigh the importance of each feature and compress them for efficiency. These designs enable our network to achieve performance comparable to state-of-the-art (SOTA) methods in real-world datasets while utilizing only a fraction of parameters, FLOPs, and runtime.
Since its introduction, UNet has been leading a variety of medical image segmentation tasks. Although numerous follow-up studies have also been dedicated to improving the performance of standard UNet, few have conducted in-depth analyses of the underlying interest pattern of UNet in medical image segmentation. In this paper, we explore the patterns learned in a UNet and observe two important factors that potentially affect its performance: (i) irrelative feature learned caused by asymmetric supervision; (ii) feature redundancy in the feature map. To this end, we propose to balance the supervision between encoder and decoder and reduce the redundant information in the UNet. Specifically, we use the feature map that contains the most semantic information (i.e., the last layer of the decoder) to provide additional supervision to other blocks to provide additional supervision and reduce feature redundancy by leveraging feature distillation. The proposed method can be easily integrated into existing UNet architecture in a plug-and-play fashion with negligible computational cost. The experimental results suggest that the proposed method consistently improves the performance of standard UNets on four medical image segmentation datasets. The code is available at \url{https://github.com/ChongQingNoSubway/SelfReg-UNet}
Multi-label classification is an essential task utilized in a wide variety of real-world applications. Multi-label zero-shot learning is a method for classifying images into multiple unseen categories for which no training data is available, while in general zero-shot situations, the test set may include observed classes. The CLIP-Decoder is a novel method based on the state-of-the-art ML-Decoder attention-based head. We introduce multi-modal representation learning in CLIP-Decoder, utilizing the text encoder to extract text features and the image encoder for image feature extraction. Furthermore, we minimize semantic mismatch by aligning image and word embeddings in the same dimension and comparing their respective representations using a combined loss, which comprises classification loss and CLIP loss. This strategy outperforms other methods and we achieve cutting-edge results on zero-shot multilabel classification tasks using CLIP-Decoder. Our method achieves an absolute increase of 3.9% in performance compared to existing methods for zero-shot learning multi-label classification tasks. Additionally, in the generalized zero-shot learning multi-label classification task, our method shows an impressive increase of almost 2.3%.
Remote sensing change detection aims to compare two or more images recorded for the same area but taken at different time stamps to quantitatively and qualitatively assess changes in geographical entities and environmental factors. Mainstream models usually built on pixel-by-pixel change detection paradigms, which cannot tolerate the diversity of changes due to complex scenes and variation in imaging conditions. To address this shortcoming, this paper rethinks the change detection with the mask view, and further proposes the corresponding: 1) meta-architecture CDMask and 2) instance network CDMaskFormer. Components of CDMask include Siamese backbone, change extractor, pixel decoder, transformer decoder and normalized detector, which ensures the proper functioning of the mask detection paradigm. Since the change query can be adaptively updated based on the bi-temporal feature content, the proposed CDMask can adapt to different latent data distributions, thus accurately identifying regions of interest changes in complex scenarios. Consequently, we further propose the instance network CDMaskFormer customized for the change detection task, which includes: (i) a Spatial-temporal convolutional attention-based instantiated change extractor to capture spatio-temporal context simultaneously with lightweight operations; and (ii) a scene-guided axial attention-instantiated transformer decoder to extract more spatial details. State-of-the-art performance of CDMaskFormer is achieved on five benchmark datasets with a satisfactory efficiency-accuracy trade-off. Code is available at https://github.com/xwmaxwma/rschange.
Self-supervised learning models have revolutionized the field of speech processing. However, the process of fine-tuning these models on downstream tasks requires substantial computational resources, particularly when dealing with multiple speech-processing tasks. In this paper, we explore the potential of adapter-based fine-tuning in developing a unified model capable of effectively handling multiple spoken language processing tasks. The tasks we investigate are Automatic Speech Recognition, Phoneme Recognition, Intent Classification, Slot Filling, and Spoken Emotion Recognition. We validate our approach through a series of experiments on the SUPERB benchmark, and our results indicate that adapter-based fine-tuning enables a single encoder-decoder model to perform multiple speech processing tasks with an average improvement of 18.4% across the five target tasks while staying efficient in terms of parameter updates.
Automated audio captioning (AAC) is an audio-to-text task to describe audio contents in natural language. Recently, the advancements in large language models (LLMs), with improvements in training approaches for audio encoders, have opened up possibilities for improving AAC. Thus, we explore enhancing AAC from three aspects: 1) a pre-trained audio encoder via consistent ensemble distillation (CED) is used to improve the effectivity of acoustic tokens, with a querying transformer (Q-Former) bridging the modality gap to LLM and compress acoustic tokens; 2) we investigate the advantages of using a Llama 2 with 7B parameters as the decoder; 3) another pre-trained LLM corrects text errors caused by insufficient training data and annotation ambiguities. Both the audio encoder and text decoder are optimized by low-rank adaptation (LoRA). Experiments show that each of these enhancements is effective. Our method obtains a 33.0 SPIDEr-FL score, outperforming the winner of DCASE 2023 Task 6A.
Even though Transformers are extensively used for Natural Language Processing tasks, especially for machine translation, they lack an explicit memory to store key concepts of processed texts. This paper explores the properties of the content of symbolic working memory added to the Transformer model decoder. Such working memory enhances the quality of model predictions in machine translation task and works as a neural-symbolic representation of information that is important for the model to make correct translations. The study of memory content revealed that translated text keywords are stored in the working memory, pointing to the relevance of memory content to the processed text. Also, the diversity of tokens and parts of speech stored in memory correlates with the complexity of the corpora for machine translation task.
Encoder--decoder neural networks (EDNN) condense information most relevant to the output of the feedforward network to activation values at a bottleneck layer. We study the use of this architecture in emulation and interpretation of simulated X-ray spectroscopic data with the aim to identify key structural characteristics for the spectra, previously studied using emulator-based component analysis (ECA). We find an EDNN to outperform ECA in covered target variable variance, but also discover complications in interpreting the latent variables in physical terms. As a compromise of the benefits of these two approaches, we develop a network where the linear projection of ECA is used, thus maintaining the beneficial characteristics of vector expansion from the latent variables for their interpretation. These results underline the necessity of information recovery after its condensation and identification of decisive structural degrees of freedom for the output spectra for a justified interpretation.
Recently, there has been increasing interest in multimodal applications that integrate text with other modalities, such as images, audio and video, to facilitate natural language interactions with multimodal AI systems. While applications involving standard modalities have been extensively explored, there is still a lack of investigation into specific data modalities such as remote sensing (RS) data. Despite the numerous potential applications of RS data, including environmental protection, disaster monitoring and land planning, available solutions are predominantly focused on specific tasks like classification, captioning and retrieval. These solutions often overlook the unique characteristics of RS data, such as its capability to systematically provide information on the same geographical areas over time. This ability enables continuous monitoring of changes in the underlying landscape. To address this gap, we propose a novel foundation model for bi-temporal RS image pairs, in the context of change detection analysis, leveraging Contrastive Learning and the LEVIR-CC dataset for both captioning and text-image retrieval. By jointly training a contrastive encoder and captioning decoder, our model add text-image retrieval capabilities, in the context of bi-temporal change detection, while maintaining captioning performances that are comparable to the state of the art. We release the source code and pretrained weights at: https://github.com/rogerferrod/RSICRC.
Semantic segmentation is an important computer vision task, particularly for scene understanding and navigation of autonomous vehicles and UAVs. Several variations of deep neural network architectures have been designed to tackle this task. However, due to their huge computational costs and their high memory consumption, these models are not meant to be deployed on resource-constrained systems. To address this limitation, we introduce an end-to-end biologically inspired semantic segmentation approach by combining Spiking Neural Networks (SNNs, a low-power alternative to classical neural networks) with event cameras whose output data can directly feed these neural network inputs. We have designed EvSegSNN, a biologically plausible encoder-decoder U-shaped architecture relying on Parametric Leaky Integrate and Fire neurons in an objective to trade-off resource usage against performance. The experiments conducted on DDD17 demonstrate that EvSegSNN outperforms the closest state-of-the-art model in terms of MIoU while reducing the number of parameters by a factor of $1.6$ and sparing a batch normalization stage.
This report introduces the first-place winning solution for the Autonomous Grand Challenge 2024 - Mapless Driving. In this report, we introduce a novel online mapping pipeline LGmap, which adept at long-range temporal model. Firstly, we propose symmetric view transformation(SVT), a hybrid view transformation module. Our approach overcomes the limitations of forward sparse feature representation and utilizing depth perception and SD prior information. Secondly, we propose hierarchical temporal fusion(HTF) module. It employs temporal information from local to global, which empowers the construction of long-range HD map with high stability. Lastly, we propose a novel ped-crossing resampling. The simplified ped crossing representation accelerates the instance attention based decoder convergence performance. Our method achieves 0.66 UniScore in the Mapless Driving OpenLaneV2 test set.
Mitigating bias in language models (LMs) has become a critical problem due to the widespread deployment of LMs. Numerous approaches revolve around data pre-processing and fine-tuning of language models, tasks that can be both time-consuming and computationally demanding. Consequently, there is a growing interest in machine unlearning techniques given their capacity to induce the forgetting of undesired behaviors of the existing pre-trained or fine-tuned models with lower computational cost. In this work, we explore two unlearning methods, (1) Partitioned Contrastive Gradient Unlearning (PCGU) applied on decoder models and (2) Negation via Task Vector, to reduce social biases in state-of-the-art and open-source LMs such as LLaMA-2 and OPT. We also implement distributed PCGU for large models. It is empirically shown, through quantitative and qualitative analyses, that negation via Task Vector method outperforms PCGU in debiasing with minimum deterioration in performance and perplexity of the models. On LLaMA-27B, negation via Task Vector reduces the bias score by 11.8%
This paper explores the performance of encoder and decoder language models on multilingual Natural Language Understanding (NLU) tasks, with a broad focus on Germanic languages. Building upon the ScandEval benchmark, initially restricted to evaluating encoder models, we extend the evaluation framework to include decoder models. We introduce a method for evaluating decoder models on NLU tasks and apply it to the languages Danish, Swedish, Norwegian, Icelandic, Faroese, German, Dutch, and English. Through a series of experiments and analyses, we also address research questions regarding the comparative performance of encoder and decoder models, the impact of NLU task types, and the variation across language resources. Our findings reveal that encoder models can achieve significantly better NLU performance than decoder models despite having orders of magnitude fewer parameters. Additionally, we investigate the correlation between decoders and task performance via a UMAP analysis, shedding light on the unique capabilities of decoder and encoder models. This study contributes to a deeper understanding of language model paradigms in NLU tasks and provides valuable insights for model selection and evaluation in multilingual settings.
Recent advancements in autoregressive networks with linear complexity have driven significant research progress, demonstrating exceptional performance in large language models. A representative model is the Extended Long Short-Term Memory (xLSTM), which incorporates gating mechanisms and memory structures, performing comparably to Transformer architectures in long-sequence language tasks. Autoregressive networks such as xLSTM can utilize image serialization to extend their application to visual tasks such as classification and segmentation. Although existing studies have demonstrated Vision-LSTM's impressive results in image classification, its performance in image semantic segmentation remains unverified. Our study represents the first attempt to evaluate the effectiveness of Vision-LSTM in the semantic segmentation of remotely sensed images. This evaluation is based on a specifically designed encoder-decoder architecture named Seg-LSTM, and comparisons with state-of-the-art segmentation networks. Our study found that Vision-LSTM's performance in semantic segmentation was limited and generally inferior to Vision-Transformers-based and Vision-Mamba-based models in most comparative tests. Future research directions for enhancing Vision-LSTM are recommended. The source code is available from https://github.com/zhuqinfeng1999/Seg-LSTM.
The entropy bottleneck introduced by Ball\'e et al. is a common component used in many learned compression models. It encodes a transformed latent representation using a static distribution whose parameters are learned during training. However, the actual distribution of the latent data may vary wildly across different inputs. The static distribution attempts to encompass all possible input distributions, thus fitting none of them particularly well. This unfortunate phenomenon, sometimes known as the amortization gap, results in suboptimal compression. To address this issue, we propose a method that dynamically adapts the encoding distribution to match the latent data distribution for a specific input. First, our model estimates a better encoding distribution for a given input. This distribution is then compressed and transmitted as an additional side-information bitstream. Finally, the decoder reconstructs the encoding distribution and uses it to decompress the corresponding latent data. Our method achieves a Bj{\o}ntegaard-Delta (BD)-rate gain of -7.10% on the Kodak test dataset when applied to the standard fully-factorized architecture. Furthermore, considering computational complexity, the transform used by our method is an order of magnitude cheaper in terms of Multiply-Accumulate (MAC) operations compared to related side-information methods such as the scale hyperprior.
In this report, we propose three novel methods for developing a sound event detection (SED) model for the DCASE 2024 Challenge Task 4. First, we propose an auxiliary decoder attached to the final convolutional block to improve feature extraction capabilities while reducing dependency on embeddings from pre-trained large models. The proposed auxiliary decoder operates independently from the main decoder, enhancing performance of the convolutional block during the initial training stages by assigning a different weight strategy between main and auxiliary decoder losses. Next, to address the time interval issue between the DESED and MAESTRO datasets, we propose maximum probability aggregation (MPA) during the training step. The proposed MPA method enables the model's output to be aligned with soft labels of 1 s in the MAESTRO dataset. Finally, we propose a multi-channel input feature that employs various versions of logmel and MFCC features to generate time-frequency pattern. The experimental results demonstrate the efficacy of these proposed methods in a view of improving SED performance by achieving a balanced enhancement across different datasets and label types. Ultimately, this approach presents a significant step forward in developing more robust and flexible SED models
Language models often struggle with handling factual knowledge, exhibiting factual hallucination issue. This makes it vital to evaluate the models' ability to recall its parametric knowledge about facts. In this study, we introduce a knowledge probing benchmark, BELIEF(ICL), to evaluate the knowledge recall ability of both encoder- and decoder-based pre-trained language models (PLMs) from diverse perspectives. BELIEFs utilize a multi-prompt dataset to evaluate PLM's accuracy, consistency, and reliability in factual knowledge recall. To enable a more reliable evaluation with BELIEFs, we semi-automatically create MyriadLAMA, which has massively diverse prompts. We validate the effectiveness of BELIEFs in comprehensively evaluating PLM's knowledge recall ability on diverse PLMs, including recent large language models (LLMs). We then investigate key factors in memorizing and recalling facts in PLMs, such as model size, pretraining strategy and corpora, instruction-tuning process and in-context learning settings. Finally, we reveal the limitation of the prompt-based knowledge probing. The MyriadLAMA is publicized.
Electroencephalogram (EEG) signals-based motor kinematics prediction (MKP) has been an active area of research to develop brain-computer interface (BCI) systems such as exosuits, prostheses, and rehabilitation devices. However, EEG source imaging (ESI) based kinematics prediction is sparsely explored in the literature. In this study, pre-movement EEG features are utilized to predict three-dimensional (3D) hand kinematics for the grasp-and-lift motor task. A public dataset, WAY-EEG-GAL, is utilized for MKP analysis. In particular, sensor-domain (EEG data) and source-domain (ESI data) based features from the frontoparietal region are explored for MKP. Deep learning-based models are explored to achieve efficient kinematics decoding. Various time-lagged and window sizes are analyzed for hand kinematics prediction. Subsequently, intra-subject and inter-subject MKP analysis is performed to investigate the subject-specific and subject-independent motor-learning capabilities of the neural decoders. The Pearson correlation coefficient (PCC) is used as the performance metric for kinematics trajectory decoding. The rEEGNet neural decoder achieved the best performance with sensor-domain and source-domain features with a time lag and window size of 100 ms and 450 ms, respectively. The highest mean PCC values of 0.790, 0.795, and 0.637 are achieved using sensor-domain features, while 0.769, 0.777, and 0.647 are achieved using source-domain features in x, y, and z-directions, respectively. This study explores the feasibility of trajectory prediction using EEG sensor-domain and source-domain EEG features for the grasp-and-lift task. Furthermore, inter-subject trajectory estimation is performed using the proposed deep learning decoder with EEG source domain features.
End-to-end multilingual speech recognition models handle multiple languages through a single model, often incorporating language identification to automatically detect the language of incoming speech. Since the common scenario is where the language is already known, these models can perform as language-specific by using language information as prompts, which is particularly beneficial for attention-based encoder-decoder architectures. However, the Connectionist Temporal Classification (CTC) approach, which enhances recognition via joint decoding and multi-task training, does not normally incorporate language prompts due to its conditionally independent output tokens. To overcome this, we introduce an encoder prompting technique within the self-conditioned CTC framework, enabling language-specific adaptation of the CTC model in a zero-shot manner. Our method has shown to significantly reduce errors by 28% on average and by 41% on low-resource languages.
In the context of real-world applications, leveraging large language models (LLMs) for domain-specific tasks often faces two major challenges: domain-specific knowledge privacy and constrained resources. To address these issues, we propose PDSS, a privacy-preserving framework for step-by-step distillation of LLMs. PDSS works on a server-client architecture, wherein client transmits perturbed prompts to the server's LLM for rationale generation. The generated rationales are then decoded by the client and used to enrich the training of task-specific small language model(SLM) within a multi-task learning paradigm. PDSS introduces two privacy protection strategies: the Exponential Mechanism Strategy and the Encoder-Decoder Strategy, balancing prompt privacy and rationale usability. Experiments demonstrate the effectiveness of PDSS in various text generation tasks, enabling the training of task-specific SLM with enhanced performance while prioritizing data privacy protection.