Xiangdong Wang

Yiming Li, Zhifang Guo, Zhirong Ye et al. · tsinghua

In this paper, we describe in detail our system for DCASE 2022 Task4. The system combines two considerably different models: an end-to-end Sound Event Detection Transformer (SEDT) and a frame-wise model, Metric Learning and Focal Loss CNN (MLFL-CNN). The former is an event-wise model which learns event-level representations and predicts sound event categories and boundaries directly, while the latter is based on the widely adopted frame-classification scheme, under which each frame is classified into event categories and event boundaries are obtained by post-processing such as thresholding and smoothing. For SEDT, self-supervised pre-training using unlabeled data is applied, and semi-supervised learning is adopted by using an online teacher, which is updated from the student model using the Exponential Moving Average (EMA) strategy and generates reliable pseudo labels for weakly-labeled and unlabeled data. For the frame-wise model, the ICT-TOSHIBA system of DCASE 2021 Task 4 is used. Experimental results show that the hybrid system considerably outperforms either individual model and achieves psds1 of 0.420 and psds2 of 0.783 on the validation set without external data. The code is available at https://github.com/965694547/Hybrid-system-of-frame-wise-model-and-SEDT.

Yachuan Li, Zongmin Li, Xavier Soria P. et al.

The significance of multi-scale features has been gradually recognized by the edge detection community. However, the fusion of multi-scale features increases the complexity of the model, which is not friendly to practical application. In this work, we propose a Compact Twice Fusion Network (CTFN) to fully integrate multi-scale features while maintaining the compactness of the model. CTFN includes two lightweight multi-scale feature fusion modules: a Semantic Enhancement Module (SEM) that can utilize the semantic information contained in coarse-scale features to guide the learning of fine-scale features, and a Pseudo Pixel-level Weighting (PPW) module that aggregate the complementary merits of multi-scale features by assigning weights to all features. Notwithstanding all this, the interference of texture noise makes the correct classification of some pixels still a challenge. For these hard samples, we propose a novel loss function, coined Dynamic Focal Loss, which reshapes the standard cross-entropy loss and dynamically adjusts the weights to correct the distribution of hard samples. We evaluate our method on three datasets, i.e., BSDS500, NYUDv2, and BIPEDv2. Compared with state-of-the-art methods, CTFN achieves competitive accuracy with less parameters and computational cost. Apart from the backbone, CTFN requires only 0.1M additional parameters, which reduces its computation cost to just 60% of other state-of-the-art methods. The codes are available at https://github.com/Li-yachuan/CTFN-pytorch-master.

Zhifang Guo, Jianguo Mao, Rui Tao et al.

Text-based audio generation models have limitations as they cannot encompass all the information in audio, leading to restricted controllability when relying solely on text. To address this issue, we propose a novel model that enhances the controllability of existing pre-trained text-to-audio models by incorporating additional conditions including content (timestamp) and style (pitch contour and energy contour) as supplements to the text. This approach achieves fine-grained control over the temporal order, pitch, and energy of generated audio. To preserve the diversity of generation, we employ a trainable control condition encoder that is enhanced by a large language model and a trainable Fusion-Net to encode and fuse the additional conditions while keeping the weights of the pre-trained text-to-audio model frozen. Due to the lack of suitable datasets and evaluation metrics, we consolidate existing datasets into a new dataset comprising the audio and corresponding conditions and use a series of evaluation metrics to evaluate the controllability performance. Experimental results demonstrate that our model successfully achieves fine-grained control to accomplish controllable audio generation. Audio samples and our dataset are publicly available at https://conditionaudiogen.github.io/conditionaudiogen/

Hualei Wang, Jianguo Mao, Zhifang Guo et al.

Large language models reveal deep comprehension and fluent generation in the field of multi-modality. Although significant advancements have been achieved in audio multi-modality, existing methods are rarely leverage language model for sound event detection (SED). In this work, we propose an end-to-end framework for understanding audio features while simultaneously generating sound event and their temporal location. Specifically, we employ pretrained acoustic models to capture discriminative features across different categories and language models for autoregressive text generation. Conventional methods generally struggle to obtain features in pure audio domain for classification. In contrast, our framework utilizes the language model to flexibly understand abundant semantic context aligned with the acoustic representation. The experimental results showcase the effectiveness of proposed method in enhancing timestamps precision and event classification.

Changhe Chen, Xiaohao Xu, Xiangdong Wang et al.

Designing soft robots is a complex and iterative process that demands cross-disciplinary expertise in materials science, mechanics, and control, often relying on intuition and extensive experimentation. While foundation models, especially Large Language Models (LLMs), have demonstrated impressive reasoning abilities, their capacity to conduct embodied design remains largely unexplored. This paper introduces RoboCrafter-QA, a novel benchmark to evaluate whether LLMs can learn representations of soft robot designs that effectively bridge the gap between high-level task descriptions and low-level morphological and material choices. RoboCrafter-QA leverages the EvoGym simulator to generate a diverse set of soft robot design challenges, spanning robotic locomotion, manipulation, and balancing tasks. Our experiments with SOTA multi-modal LLMs reveal that while these models exhibit promising capabilities in learning design representations, they struggle with fine-grained distinctions between designs with subtle performance differences. To overcome these limitations, we finetune an efficient, open-source LLM that achieves SOTA performance on our benchmark, demonstrating superior capabilities in both design selection and direct generation of high-performing robot morphologies. Furthermore, we construct a physical replica of the modular soft robot and demonstrate a strong sim-to-real correlation, validating that superior benchmark performance has the potential to translate to effective real-world design selection. Our full system will be open-sourced to foster this exciting direction.

Zekang Yang, Hong Liu, Xiangdong Wang

Cancer survival prediction is a challenging task that involves analyzing of the tumor microenvironment within Whole Slide Image (WSI). Previous methods cannot effectively capture the intricate interaction features among instances within the local area of WSI. Moreover, existing methods for cancer survival prediction based on WSI often fail to provide better clinically meaningful predictions. To overcome these challenges, we propose a Sparse Context-aware Multiple Instance Learning (SCMIL) framework for predicting cancer survival probability distributions. SCMIL innovatively segments patches into various clusters based on their morphological features and spatial location information, subsequently leveraging sparse self-attention to discern the relationships between these patches with a context-aware perspective. Considering many patches are irrelevant to the task, we introduce a learnable patch filtering module called SoftFilter, which ensures that only interactions between task-relevant patches are considered. To enhance the clinical relevance of our prediction, we propose a register-based mixture density network to forecast the survival probability distribution for individual patients. We evaluate SCMIL on two public WSI datasets from the The Cancer Genome Atlas (TCGA) specifically focusing on lung adenocarcinom (LUAD) and kidney renal clear cell carcinoma (KIRC). Our experimental results indicate that SCMIL outperforms current state-of-the-art methods for survival prediction, offering more clinically meaningful and interpretable outcomes. Our code is accessible at https://github.com/yang-ze-kang/SCMIL.

Rui Tao, Long Yan, Kazushige Ouchi et al.

The recently proposed Mean Teacher method, which exploits large-scale unlabeled data in a self-ensembling manner, has achieved state-of-the-art results in several semi-supervised learning benchmarks. Spurred by current achievements, this paper proposes an effective Couple Learning method that combines a well-trained model and a Mean Teacher model. The suggested pseudo-labels generated model (PLG) increases strongly- and weakly-labeled data to improve the Mean Teacher method-s performance. Moreover, the Mean Teacher-s consistency cost reduces the noise impact in the pseudo-labels introduced by detection errors. The experimental results on Task 4 of the DCASE2020 challenge demonstrate the superiority of the proposed method, achieving about 44.25% F1-score on the public evaluation set, significantly outperforming the baseline system-s 32.39%. At the same time, we also propose a simple and effective experiment called the Variable Order Input (VOI) experiment, which proves the significance of the Couple Learning method. Our developed Couple Learning code is available on GitHub.

Zekang Yang, Hong Liu, Xiangdong Wang

Deep learning has been extensively researched in the analysis of pathology whole-slide images (WSIs). However, most existing methods are limited to providing prediction interpretability by locating the model's salient areas in a post-hoc manner, failing to offer more reliable and accountable explanations. In this work, we propose Phenotype-Aware Multiple Instance Learning (PA-MIL), a novel ante-hoc interpretable framework that identifies cancer-related phenotypes from WSIs and utilizes them for cancer subtyping. To facilitate PA-MIL in learning phenotype-aware features, we 1) construct a phenotype knowledge base containing cancer-related phenotypes and their associated genotypes. 2) utilize the morphological descriptions of phenotypes as language prompting to aggregate phenotype-related features. 3) devise the Genotype-to-Phenotype Neural Network (GP-NN) grounded in genotype-to-phenotype relationships, which provides multi-level guidance for PA-MIL. Experimental results on multiple datasets demonstrate that PA-MIL achieves competitive performance compared to existing MIL methods while offering improved interpretability. PA-MIL leverages phenotype saliency as evidence and, using a linear classifier, achieves competitive results compared to state-of-the-art methods. Additionally, we thoroughly analyze the genotype-phenotype relationships, as well as cohort-level and case-level interpretability, demonstrating the reliability and accountability of PA-MIL.

Zekang Yang, Hong Liu, Xiangdong Wang

Computer-aided cancer survival risk prediction plays an important role in the timely treatment of patients. This is a challenging weakly supervised ordinal regression task associated with multiple clinical factors involved such as pathological images, genomic data and etc. In this paper, we propose a new training method, multimodal object-level contrast learning, for cancer survival risk prediction. First, we construct contrast learning pairs based on the survival risk relationship among the samples in the training sample set. Then we introduce the object-level contrast learning method to train the survival risk predictor. We further extend it to the multimodal scenario by applying cross-modal constrast. Considering the heterogeneity of pathological images and genomics data, we construct a multimodal survival risk predictor employing attention-based and self-normalizing based nerural network respectively. Finally, the survival risk predictor trained by our proposed method outperforms state-of-the-art methods on two public multimodal cancer datasets for survival risk prediction.

Zhirong Ye, Xiangdong Wang, Hong Liu et al.

Recently, an event-based end-to-end model (SEDT) has been proposed for sound event detection (SED) and achieves competitive performance. However, compared with the frame-based model, it requires more training data with temporal annotations to improve the localization ability. Synthetic data is an alternative, but it suffers from a great domain gap with real recordings. Inspired by the great success of UP-DETR in object detection, we propose to self-supervisedly pre-train SEDT (SP-SEDT) by detecting random patches (only cropped along the time axis). Experiments on the DCASE2019 task4 dataset show the proposed SP-SEDT can outperform fine-tuned frame-based model. The ablation study is also conducted to investigate the impact of different loss functions and patch size.

Zhirong Ye, Xiangdong Wang, Hong Liu et al.

Sound event detection (SED) has gained increasing attention with its wide application in surveillance, video indexing, etc. Existing models in SED mainly generate frame-level prediction, converting it into a sequence multi-label classification problem. A critical issue with the frame-based model is that it pursues the best frame-level prediction rather than the best event-level prediction. Besides, it needs post-processing and cannot be trained in an end-to-end way. This paper firstly presents the one-dimensional Detection Transformer (1D-DETR), inspired by Detection Transformer for image object detection. Furthermore, given the characteristics of SED, the audio query branch and a one-to-many matching strategy for fine-tuning the model are added to 1D-DETR to form Sound Event Detection Transformer (SEDT). To our knowledge, SEDT is the first event-based and end-to-end SED model. Experiments are conducted on the URBAN-SED dataset and the DCASE2019 Task4 dataset, and both show that SEDT can achieve competitive performance.

Yuxin Huang, Liwei Lin, Xiangdong Wang et al.

Due to the limitation of strong-labeled sound event detection data set, using synthetic data to improve the sound event detection system performance has been a new research focus. In this paper, we try to exploit the usage of synthetic data to improve the feature representation. Based on metric learning, we proposed inter-frame distance loss function for domain adaptation, and prove the effectiveness of it on sound event detection. We also applied multi-task learning with synthetic data. We find the the best performance can be achieved when the two methods being used together. The experiment on DCASE 2018 task 4 test set and DCASE 2019 task 4 synthetic set both show competitive results.

Yuxin Huang, Liwei Lin, Shuo Ma et al.

In this paper, we describe in detail our systems for DCASE 2020 Task 4. The systems are based on the 1st-place system of DCASE 2019 Task 4, which adopts weakly-supervised framework with an attention-based embedding-level pooling module and a semi-supervised learning approach named guided learning. This year, we incorporate multi-branch learning (MBL) into the original system to further improve its performance. MBL uses different branches with different pooling strategies (including instance-level and embedding-level strategies) and different pooling modules (including attention pooling, global max pooling or global average pooling modules), which share the same feature encoder of the model. Therefore, multiple branches pursuing different purposes and focusing on different characteristics of the data can help the feature encoder model the feature space better and avoid over-fitting. To better exploit the strongly-labeled synthetic data, inspired by multi-task learning, we also employ a sound event detection branch. To combine sound separation (SS) with sound event detection (SED), we fuse the results of SED systems with SS-SED systems which are trained using separated sound output by an SS system. The experimental results prove that MBL can improve the model performance and using SS has great potential to improve the performance of SED ensemble system.

Yong Ruan, Xiangdong Wang, Hong Liu et al.

The dominant automatic lexical stress detection method is to split the utterance into syllable segments using phoneme sequence and their time-aligned boundaries. Then we extract features from syllable to use classification method to classify the lexical stress. However, we can't get very accurate time boundaries of each phoneme and we have to design some features in the syllable segments to classify the lexical stress. Therefore, we propose a end-to-end approach using sequence to sequence model of transformer to estimate lexical stress. For this, we train transformer model using feature sequence of audio and their phoneme sequence with lexical stress marks. During the recognition process, the recognized phoneme sequence is restricted according to the original standard phoneme sequence without lexical stress marks, but the lexical stress mark of each phoneme is not limited. We train the model in different subset of Librispeech and do lexical stress recognition in TIMIT and L2-ARCTIC dataset. For all subsets, the end-to-end model will perform better than the syllable segments classification method. Our method can achieve a 6.36% phoneme error rate on the TIMIT dataset, which exceeds the 7.2% error rate in other studies.

Liwei Lin, Xiangdong Wang, Hong Liu et al.

In this paper, we describe in detail the system we submitted to DCASE2019 task 4: sound event detection (SED) in domestic environments. We employ a convolutional neural network (CNN) with an embedding-level attention pooling module to solve it. By considering the interference caused by the co-occurrence of multiple events in the unbalanced dataset, we utilize the disentangled feature to raise the performance of the model. To take advantage of the unlabeled data, we adopt Guided Learning for semi-supervised learning. A group of median filters with adaptive window sizes is utilized in the post-processing of output probabilities of the model. We also analyze the effect of the synthetic data on the performance of the model and finally achieve an event-based F-measure of 45.43% on the validation set and an event-based F-measure of 42.7% on the test set. The system we submitted to the challenge achieves the best performance compared to those of other participates.

Liwei Lin, Xiangdong Wang, Hong Liu et al.

We propose a simple but efficient method termed Guided Learning for weakly-labeled semi-supervised sound event detection (SED). There are two sub-targets implied in weakly-labeled SED: audio tagging and boundary detection. Instead of designing a single model by considering a trade-off between the two sub-targets, we design a teacher model aiming at audio tagging to guide a student model aiming at boundary detection to learn using the unlabeled data. The guidance is guaranteed by the audio tagging performance gap of the two models. In the meantime, the student model liberated from the trade-off is able to provide more excellent boundary detection results. We propose a principle to design such two models based on the relation between the temporal compression scale and the two sub-targets. We also propose an end-to-end semi-supervised learning process for these two models to enable their abilities to rise alternately. Experiments on the DCASE2018 Task4 dataset show that our approach achieves competitive performance.

Liwei Lin, Xiangdong Wang, Hong Liu et al.

In this paper, a special decision surface for the weakly-supervised sound event detection (SED) and a disentangled feature (DF) for the multi-label problem in polyphonic SED are proposed. We approach SED as a multiple instance learning (MIL) problem and utilize a neural network framework with a pooling module to solve it. General MIL approaches include two kinds: the instance-level approaches and embedding-level approaches. We present a method of generating instance-level probabilities for the embedding level approaches which tend to perform better than the instance-level approaches in terms of bag-level classification but can not provide instance-level probabilities in current approaches. Moreover, we further propose a specialized decision surface (SDS) for the embedding-level attention pooling. We analyze and explained why an embedding-level attention module with SDS is better than other typical pooling modules from the perspective of the high-level feature space. As for the problem of the unbalanced dataset and the co-occurrence of multiple categories in the polyphonic event detection task, we propose a DF to reduce interference among categories, which optimizes the high-level feature space by disentangling it based on class-wise identifiable information and obtaining multiple different subspaces. Experiments on the dataset of DCASE 2018 Task 4 show that the proposed SDS and DF significantly improve the detection performance of the embedding-level MIL approach with an attention pooling module and outperform the first place system in the challenge by 6.6 percentage points.