arXiv subjects

Haoqi Zhu

Publications and source records attributed to Haoqi Zhu.

7 recordsLinked to original sources

arXiv2026

SMM Transformer: Leveraging Spiking Neural Networks for Multimodal Tasks

Spiking Neural Networks (SNNs) enable event-driven computation with sparse activations, but building multimodal Transformers on SNNs is hindered by unstable training in deep spiking stacks and the mismatch between dense softmax attention and spike-based communication. We propose SMM Transformer, an SNN-based multimodal Transformer framework that combines (i)PLMP, a Parallel LIF with Multistage Learnable Parameters neuron and a tailored P-STBP algorithm for stable deep SNN training, (ii) SMSA, an attention-inspired spike-driven token-mixing module that replaces dense pairwise softmax attention with channel-wise spike co-activation and self-compensation, and (iii)SMoE, a spiking mixture-of-experts module for modality-aware fusion. Across visual and multimodal benchmarks, SMM Transformer achieves competitive accuracy compared to ANN baselines. Under a standard MAC/AC arithmetic model, SMSA reduces the estimated operator-level compute energy of the attention module by up to 97%, while whole-model profiling shows more moderate but consistent efficiency gains.

cs.NE↗

arXiv2026

Spike-HTR: Spiking Neural Transformer for Handwritten Text Recognition

Handwritten Text Recognition (HTR) is computationally imbalanced in two ways: most image pixels are background, and many width-axis sequence positions are blank-dominated. This creates a mismatch for Spiking Neural Networks (SNNs): handwriting is observed as a static image, whereas spiking computation unfolds over timesteps. We propose Spike-HTR, a hybrid spiking recognizer that controls both the number of spiking steps and the number of width positions processed by the deep sequence mixer. To make a static image suitable for short-horizon spiking inference, InkCoder converts it into a coarse-to-fine input stream, where early steps cover broad stroke regions and later steps emphasize sharper stroke details. To reduce sequence computation, a CTC-guided length reducer keeps likely character or uncertain positions and compresses long blank-dominated stretches before deep mixing. With $T{=}2$, Spike-HTR trains only on target data, decodes without language models or lexicons, and reaches validation/test CERs of 3.5/5.4, 2.3/2.5, and 4.2/3.9 on IAM, LAM, and READ2016. Codes are available at https://github.com/QomolangmaH/SpikeHTR.

cs.NE↗

arXiv2025

Surface mobility of a glass-forming polymer in an ionic liquid

The free surface of glassy polymers exhibits enhanced segmental dynamics compared to the bulk, forming a liquid-like layer that lowers the glass transition temperature (Tg) in nanometersized polymer samples. Recent studies have shown that immersing polymers in ionic liquids can suppress this enhanced surface dynamics. To investigate how ionic liquids influence polymer dynamics near the ionic-liquid-polymer interface, we measure the surface leveling of nanometersized stepped polystyrene films immersed in ionic liquids, and compared the results to the case of films in vacuum. Our results reveal that ionic liquids significantly slow the leveling process both above and below Tg. However, our results indicate that the liquid-like surface layer below Tg does exist in ionic liquids. Numerical solutions of the thin-film equation, incorporating appropriate boundary conditions, show that the surface mobility of PS films in ionic liquids can match that of PS films in vacuum. Thus, while ionic liquids alter the polymer flow process, they do not eliminate the dynamical heterogeneity inherent to glassy polymers.

cond-mat.soft↗

arXiv2024

Cross-Modal Denoising: A Novel Training Paradigm for Enhancing Speech-Image Retrieval

The success of speech-image retrieval relies on establishing an effective alignment between speech and image. Existing methods often model cross-modal interaction through simple cosine similarity of the global feature of each modality, which fall short in capturing fine-grained details within modalities. To address this issue, we introduce an effective framework and a novel learning task named cross-modal denoising (CMD) to enhance cross-modal interaction to achieve finer-level cross-modal alignment. Specifically, CMD is a denoising task designed to reconstruct semantic features from noisy features within one modality by interacting features from another modality. Notably, CMD operates exclusively during model training and can be removed during inference without adding extra inference time. The experimental results demonstrate that our framework outperforms the state-of-the-art method by 2.0% in mean R@1 on the Flickr8k dataset and by 1.7% in mean R@1 on the SpokenCOCO dataset for the speech-image retrieval tasks, respectively. These experimental results validate the efficiency and effectiveness of our framework.

cs.CL↗

arXiv2024

HaltingVT: Adaptive Token Halting Transformer for Efficient Video Recognition

Action recognition in videos poses a challenge due to its high computational cost, especially for Joint Space-Time video transformers (Joint VT). Despite their effectiveness, the excessive number of tokens in such architectures significantly limits their efficiency. In this paper, we propose HaltingVT, an efficient video transformer adaptively removing redundant video patch tokens, which is primarily composed of a Joint VT and a Glimpser module. Specifically, HaltingVT applies data-adaptive token reduction at each layer, resulting in a significant reduction in the overall computational cost. Besides, the Glimpser module quickly removes redundant tokens in shallow transformer layers, which may even be misleading for video recognition tasks based on our observations. To further encourage HaltingVT to focus on the key motion-related information in videos, we design an effective Motion Loss during training. HaltingVT acquires video analysis capabilities and token halting compression strategies simultaneously in a unified training process, without requiring additional training procedures or sub-networks. On the Mini-Kinetics dataset, we achieved 75.0% top-1 ACC with 24.2 GFLOPs, as well as 67.2% top-1 ACC with an extremely low 9.9 GFLOPs. The code is available at https://github.com/dun-research/HaltingVT.

cs.CV↗

arXiv2022

Acoustic Pornography Recognition Using Convolutional Neural Networks and Bag of Refinements

A large number of pornographic audios publicly available on the Internet seriously threaten the mental and physical health of children, but these audios are rarely detected and filtered. In this paper, we firstly propose a convolutional neural networks (CNN) based model for acoustic pornography recognition. Then, we research a collection of refinements and verify their effectiveness through ablation studies. Finally, we stack all refinements together to verify whether they can further improve the accuracy of the model. Experimental results on our newly-collected large dataset consisting of 224127 pornographic audios and 274206 normal samples demonstrate the effectiveness of our proposed model and these refinements. Specifically, the proposed model achieves an accuracy of 92.46% and the accuracy is further improved to 97.19% when all refinements are combined.

cs.SD↗

arXiv2019

Creating Lightweight Object Detectors with Model Compression for Deployment on Edge Devices

To achieve lightweight object detectors for deployment on the edge devices, an effective model compression pipeline is proposed in this paper. The compression pipeline consists of automatic channel pruning for the backbone, fixed channel deletion for the branch layers and knowledge distillation for the guidance learning. As results, the Resnet50-v1d is auto-pruned and fine-tuned on ImageNet to attain a compact base model as the backbone of object detector. Then, lightweight object detectors are implemented with proposed compression pipeline. For instance, the SSD-300 with model size=16.3MB, FLOPS=2.31G, and mAP=71.2 is created, revealing a better result than SSD-300-MobileNet.

cs.CV↗