The X-LANCE Technical Report for Interspeech 2024 Speech Processing Using Discrete Speech Unit Challenge

Discrete speech tokens have been more and more popular in multiple speech processing fields, including automatic speech recognition (ASR), text-to-speech (TTS) and singing voice synthesis (SVS). In this paper, we describe the systems developed by the SJTU X-LANCE group for the TTS (acoustic + vocoder), SVS, and ASR tracks in the Interspeech 2024 Speech Processing Using Discrete Speech Unit Challenge. Notably, we achieved 1st rank on the leaderboard in the TTS track both with the whole training set and only 1h training data, with the highest UTMOS score and lowest bitrate among all submissions.

H2RSVLM: Towards Helpful and Honest Remote Sensing Large Vision Language Model

The generic large Vision-Language Models (VLMs) is rapidly developing, but still perform poorly in Remote Sensing (RS) domain, which is due to the unique and specialized nature of RS imagery and the comparatively limited spatial perception of current VLMs. Existing Remote Sensing specific Vision Language Models (RSVLMs) still have considerable potential for improvement, primarily owing to the lack of large-scale, high-quality RS vision-language datasets. We constructed HqDC-1.4M, the large scale High quality and Detailed Captions for RS images, containing 1.4 million image-caption pairs, which not only enhance the RSVLM's understanding of RS images but also significantly improve the model's spatial perception abilities, such as localization and counting, thereby increasing the helpfulness of the RSVLM. Moreover, to address the inevitable "hallucination" problem in RSVLM, we developed RSSA, the first dataset aimed at enhancing the Self-Awareness capability of RSVLMs. By incorporating a variety of unanswerable questions into typical RS visual question-answering tasks, RSSA effectively improves the truthfulness and reduces the hallucinations of the model's outputs, thereby enhancing the honesty of the RSVLM. Based on these datasets, we proposed the H2RSVLM, the Helpful and Honest Remote Sensing Vision Language Model. H2RSVLM has achieved outstanding performance on multiple RS public datasets and is capable of recognizing and refusing to answer the unanswerable questions, effectively mitigating the incorrect generations. We will release the code, data and model weights at https://github.com/opendatalab/H2RSVLM .

Information Cascade Prediction under Public Emergencies: A Survey

With the advent of the era of big data, massive information, expert experience, and high-accuracy models bring great opportunities to the information cascade prediction of public emergencies. However, the involvement of specialist knowledge from various disciplines has resulted in a primarily application-specific focus (e.g., earthquakes, floods, infectious diseases) for information cascade prediction of public emergencies. The lack of a unified prediction framework poses a challenge for classifying intersectional prediction methods across different application fields. This survey paper offers a systematic classification and summary of information cascade modeling, prediction, and application. We aim to help researchers identify cutting-edge research and comprehend models and methods of information cascade prediction under public emergencies. By summarizing open issues and outlining future directions in this field, this paper has the potential to be a valuable resource for researchers conducting further studies on predicting information cascades.

Graph Image Prior for Unsupervised Dynamic MRI Reconstruction

The inductive bias of the convolutional neural network (CNN) can act as a strong prior for image restoration, which is known as the Deep Image Prior (DIP). In recent years, DIP has been utilized in unsupervised dynamic MRI reconstruction, which adopts a generative model from the latent space to the image space. However, existing methods usually utilize a single pyramid-shaped CNN architecture to parameterize the generator, which cannot effectively exploit the spatio-temporal correlations within the dynamic data. In this work, we propose a novel scheme to exploit the DIP prior for dynamic MRI reconstruction, named ``Graph Image Prior'' (GIP). The generative model is decomposed into two stages: image recovery and manifold discovery, which is bridged by a graph convolutional network to exploit the spatio-temporal correlations. In addition, we devise an ADMM algorithm to alternately optimize the images and the network parameters to further improve the reconstruction performance. Experimental results demonstrate that GIP outperforms compressed sensing methods and unsupervised methods over different sampling trajectories, and significantly reduces the performance gap with the state-of-art supervised deep-learning methods. Moreover, GIP displays superior generalization ability when transferred to a different reconstruction setting, without the need for any additional data.

Safeguarding Medical Image Segmentation Datasets against Unauthorized Training via Contour- and Texture-Aware Perturbations

The widespread availability of publicly accessible medical images has significantly propelled advancements in various research and clinical fields. Nonetheless, concerns regarding unauthorized training of AI systems for commercial purposes and the duties of patient privacy protection have led numerous institutions to hesitate to share their images. This is particularly true for medical image segmentation (MIS) datasets, where the processes of collection and fine-grained annotation are time-intensive and laborious. Recently, Unlearnable Examples (UEs) methods have shown the potential to protect images by adding invisible shortcuts. These shortcuts can prevent unauthorized deep neural networks from generalizing. However, existing UEs are designed for natural image classification and fail to protect MIS datasets imperceptibly as their protective perturbations are less learnable than important prior knowledge in MIS, e.g., contour and texture features. To this end, we propose an Unlearnable Medical image generation method, termed UMed. UMed integrates the prior knowledge of MIS by injecting contour- and texture-aware perturbations to protect images. Given that our target is to only poison features critical to MIS, UMed requires only minimal perturbations within the ROI and its contour to achieve greater imperceptibility (average PSNR is 50.03) and protective performance (clean average DSC degrades from 82.18% to 6.80%).

NN-Defined Modulator: Reconfigurable and Portable Software Modulator on IoT Gateways

A physical-layer modulator is a vital component for an IoT gateway to map the symbols to signals. However, due to the soldered hardware chipsets on the gateway's motherboards or the diverse toolkits on different platforms for the software radio, the existing solutions either have limited extensibility or are platform-specific. Such limitation is hard to ignore when modulation schemes and hardware platforms have become extremely diverse. This paper presents a new paradigm of using neural networks as an abstraction layer for physical layer modulators in IoT gateway devices, referred to as NN-defined modulators. Our approach addresses the challenges of extensibility and portability for multiple technologies on various hardware platforms. The proposed NN-defined modulator uses a model-driven methodology rooted in solid mathematical foundations while having native support for hardware acceleration and portability to heterogeneous platforms. We conduct the evaluation of NN-defined modulators on different platforms, including Nvidia Jetson Nano and Raspberry Pi. Evaluations demonstrate that our NN-defined modulator effectively operates as conventional modulators and provides significant efficiency gains (up to $4.7\times$ on Nvidia Jetson Nano and $1.1\times$ on Raspberry Pi), indicating high portability. Furthermore, we show the real-world applications using our NN-defined modulators to generate ZigBee and WiFi packets, which are compliant with commodity TI CC2650 (ZigBee) and Intel AX201 (WiFi NIC), respectively.

NeuPAN: Direct Point Robot Navigation with End-to-End Model-based Learning

Navigating a nonholonomic robot in a cluttered environment requires extremely accurate perception and locomotion for collision avoidance. This paper presents NeuPAN: a real-time, highly-accurate, map-free, robot-agnostic, and environment-invariant robot navigation solution. Leveraging a tightly-coupled perception-locomotion framework, NeuPAN has two key innovations compared to existing approaches: 1) it directly maps raw points to a learned multi-frame distance space, avoiding error propagation from perception to control; 2) it is interpretable from an end-to-end model-based learning perspective, enabling provable convergence. The crux of NeuPAN is to solve a high-dimensional end-to-end mathematical model with various point-level constraints using the plug-and-play (PnP) proximal alternating-minimization network (PAN) with neurons in the loop. This allows NeuPAN to generate real-time, end-to-end, physically-interpretable motions directly from point clouds, which seamlessly integrates data- and knowledge-engines, where its network parameters are adjusted via back propagation. We evaluate NeuPAN on car-like robot, wheel-legged robot, and passenger autonomous vehicle, in both simulated and real-world environments. Experiments demonstrate that NeuPAN outperforms various benchmarks, in terms of accuracy, efficiency, robustness, and generalization capability across various environments, including the cluttered sandbox, office, corridor, and parking lot. We show that NeuPAN works well in unstructured environments with arbitrary-shape undetectable objects, making impassable ways passable.

From Instructions to Constraints: Language Model Alignment with Automatic Constraint Verification

User alignment is crucial for adapting general-purpose language models (LMs) to downstream tasks, but human annotations are often not available for all types of instructions, especially those with customized constraints. We observe that user instructions typically contain constraints. While assessing response quality in terms of the whole instruction is often costly, efficiently evaluating the satisfaction rate of constraints is feasible. We investigate common constraints in NLP tasks, categorize them into three classes based on the types of their arguments, and propose a unified framework, ACT (Aligning to ConsTraints), to automatically produce supervision signals for user alignment with constraints. Specifically, ACT uses constraint verifiers, which are typically easy to implement in practice, to compute constraint satisfaction rate (CSR) of each response. It samples multiple responses for each prompt and collect preference labels based on their CSR automatically. Subsequently, ACT adapts the LM to the target task through a ranking-based learning process. Experiments on fine-grained entity typing, abstractive summarization, and temporal question answering show that ACT is able to enhance LMs' capability to adhere to different classes of constraints, thereby improving task performance. Further experiments show that the constraint-following capabilities are transferable.

Seamless Virtual Reality with Integrated Synchronizer and Synthesizer for Autonomous Driving

Virtual reality (VR) is a promising data engine for autonomous driving (AD). However, data fidelity in this paradigm is often degraded by VR inconsistency, for which the existing VR approaches become ineffective, as they ignore the inter-dependency between low-level VR synchronizer designs (i.e., data collector) and high-level VR synthesizer designs (i.e., data processor). This paper presents a seamless virtual reality SVR platform for AD, which mitigates such inconsistency, enabling VR agents to interact with each other in a shared symbiotic world. The crux to SVR is an integrated synchronizer and synthesizer IS2 design, which consists of a drift-aware lidar-inertial synchronizer for VR colocation and a motion-aware deep visual synthesis network for augmented reality image generation. We implement SVR on car-like robots in two sandbox platforms, achieving a cm-level VR colocalization accuracy and 3.2% VR image deviation, thereby avoiding missed collisions or model clippings. Experiments show that the proposed SVR reduces the intervention times, missed turns, and failure rates compared to other benchmarks. The SVR-trained neural network can handle unseen situations in real-world environments, by leveraging its knowledge learnt from the VR space.

Suppress and Rebalance: Towards Generalized Multi-Modal Face Anti-Spoofing

Face Anti-Spoofing (FAS) is crucial for securing face recognition systems against presentation attacks. With advancements in sensor manufacture and multi-modal learning techniques, many multi-modal FAS approaches have emerged. However, they face challenges in generalizing to unseen attacks and deployment conditions. These challenges arise from (1) modality unreliability, where some modality sensors like depth and infrared undergo significant domain shifts in varying environments, leading to the spread of unreliable information during cross-modal feature fusion, and (2) modality imbalance, where training overly relies on a dominant modality hinders the convergence of others, reducing effectiveness against attack types that are indistinguishable sorely using the dominant modality. To address modality unreliability, we propose the Uncertainty-Guided Cross-Adapter (U-Adapter) to recognize unreliably detected regions within each modality and suppress the impact of unreliable regions on other modalities. For modality imbalance, we propose a Rebalanced Modality Gradient Modulation (ReGrad) strategy to rebalance the convergence speed of all modalities by adaptively adjusting their gradients. Besides, we provide the first large-scale benchmark for evaluating multi-modal FAS performance under domain generalization scenarios. Extensive experiments demonstrate that our method outperforms state-of-the-art methods. Source code and protocols will be released on https://github.com/OMGGGGG/mmdg.