Wei Chen

Jiahui Yang, Donglin Di, Baorui Ma et al.

In recent years, advancements in generative models have significantly expanded the capabilities of text-to-3D generation. Many approaches rely on Score Distillation Sampling (SDS) technology. However, SDS struggles to accommodate multi-condition inputs, such as text and visual prompts, in customized generation tasks. To explore the core reasons, we decompose SDS into a difference term and a classifier-free guidance term. Our analysis identifies the core issue as arising from the difference term and the random noise addition during the optimization process, both contributing to deviations from the target mode during distillation. To address this, we propose a novel algorithm, Classifier Score Matching (CSM), which removes the difference term in SDS and uses a deterministic noise addition process to reduce noise during optimization, effectively overcoming the low-quality limitations of SDS in our customized generation framework. Based on CSM, we integrate visual prompt information with an attention fusion mechanism and sampling guidance techniques, forming the Visual Prompt CSM (VPCSM) algorithm. Furthermore, we introduce a Semantic-Geometry Calibration (SGC) module to enhance quality through improved textual information integration. We present our approach as TV-3DG, with extensive experiments demonstrating its capability to achieve stable, high-quality, customized 3D generation. Project page: \url{https://yjhboy.github.io/TV-3DG}

Wei Chen, Binzhu Sha, Jing Yang et al.

Melody preservation is crucial in singing voice conversion (SVC). However, in many scenarios, audio is often accompanied with background music (BGM), which can cause audio distortion and interfere with the extraction of melody and other key features, significantly degrading SVC performance. Previous methods have attempted to address this by using more robust neural network-based melody extractors, but their performance drops sharply in the presence of complex accompaniment. Other approaches involve performing source separation before conversion, but this often introduces noticeable artifacts, leading to a significant drop in conversion quality and increasing the user's operational costs. To address these issues, we introduce a novel SVC method that uses self-supervised representation-based melody features to improve melody modeling accuracy in the presence of BGM. In our experiments, we compare the effectiveness of different self-supervised learning (SSL) models for melody extraction and explore for the first time how SSL benefits the task of melody extraction. The experimental results demonstrate that our proposed SVC model significantly outperforms existing baseline methods in terms of melody accuracy and shows higher similarity and naturalness in both subjective and objective evaluations across noisy and clean audio environments.

Kartikeya Bhardwaj, Wei Chen, Radu Marculescu

In this paper, we first highlight three major challenges to large-scale adoption of deep learning at the edge: (i) Hardware-constrained IoT devices, (ii) Data security and privacy in the IoT era, and (iii) Lack of network-aware deep learning algorithms for distributed inference across multiple IoT devices. We then provide a unified view targeting three research directions that naturally emerge from the above challenges: (1) Federated learning for training deep networks, (2) Data-independent deployment of learning algorithms, and (3) Communication-aware distributed inference. We believe that the above research directions need a network-centric approach to enable the edge intelligence and, therefore, fully exploit the true potential of IoT.

Wei Chen, Jian Sun, Weishuo Li et al.

Obstacle avoidance is one of the essential and indispensable functions for autonomous mobile robots. Most of the existing solutions are typically based on single condition constraint and cannot incorporate sensor data in a real-time manner, which often fail to respond to unexpected moving obstacles in dynamic unknown environments. In this paper, a novel real-time multi-constraints obstacle avoidance method using Light Detection and Ranging(LiDAR) is proposed, which is able to, based on the latest estimation of the robot pose and environment, find the sub-goal defined by a multi-constraints function within the explored region and plan a corresponding optimal trajectory at each time step iteratively, so that the robot approaches the goal over time. Meanwhile, at each time step, the improved Ant Colony Optimization(ACO) algorithm is also used to re-plan optimal paths from the latest robot pose to the latest defined sub-goal position. While ensuring convergence, planning in this method is done by repeated local optimizations, so that the latest sensor data from LiDAR and derived environment information can be fully utilized at each step until the robot reaches the desired position. This method facilitates real-time performance, also has little requirement on memory space or computational power due to its nature, thus our method has huge potentials to benefit small low-cost autonomous platforms. The method is evaluated against several existing technologies in both simulation and real-world experiments.

Wei Chen, Kartikeya Bhardwaj, Radu Marculescu

In this paper, we identify a new phenomenon called activation-divergence which occurs in Federated Learning (FL) due to data heterogeneity (i.e., data being non-IID) across multiple users. Specifically, we argue that the activation vectors in FL can diverge, even if subsets of users share a few common classes with data residing on different devices. To address the activation-divergence issue, we introduce a prior based on the principle of maximum entropy; this prior assumes minimal information about the per-device activation vectors and aims at making the activation vectors of same classes as similar as possible across multiple devices. Our results show that, for both IID and non-IID settings, our proposed approach results in better accuracy (due to the significantly more similar activation vectors across multiple devices), and is more communication-efficient than state-of-the-art approaches in FL. Finally, we illustrate the effectiveness of our approach on a few common benchmarks and two large medical datasets.

Ruchao Fan, Pan Zhou, Wei Chen et al.

Attention-based end-to-end models such as Listen, Attend and Spell (LAS), simplify the whole pipeline of traditional automatic speech recognition (ASR) systems and become popular in the field of speech recognition. In previous work, researchers have shown that such architectures can acquire comparable results to state-of-the-art ASR systems, especially when using a bidirectional encoder and global soft attention (GSA) mechanism. However, bidirectional encoder and GSA are two obstacles for real-time speech recognition. In this work, we aim to stream LAS baseline by removing the above two obstacles. On the encoder side, we use a latency-controlled (LC) bidirectional structure to reduce the delay of forward computation. Meanwhile, an adaptive monotonic chunk-wise attention (AMoChA) mechanism is proposed to replace GSA for the calculation of attention weight distribution. Furthermore, we propose two methods to alleviate the huge performance degradation when combining LC and AMoChA. Finally, we successfully acquire an online LAS model, LC-AMoChA, which has only 3.5% relative performance reduction to LAS baseline on our internal Mandarin corpus.