arXiv subjects

Tao Fang

Publications and source records attributed to Tao Fang.

At least 37 records · Page 2Linked to original sources

arXiv2025

MoTAS: MoE-Guided Feature Selection from TTS-Augmented Speech for Enhanced Multimodal Alzheimer's Early Screening

Early screening for Alzheimer's Disease (AD) through speech presents a promising non-invasive approach. However, challenges such as limited data and the lack of fine-grained, adaptive feature selection often hinder performance. To address these issues, we propose MoTAS, a robust framework designed to enhance AD screening efficiency. MoTAS leverages Text-to-Speech (TTS) augmentation to increase data volume and employs a Mixture of Experts (MoE) mechanism to improve multimodal feature selection, jointly enhancing model generalization. The process begins with automatic speech recognition (ASR) to obtain accurate transcriptions. TTS is then used to synthesize speech that enriches the dataset. After extracting acoustic and text embeddings, the MoE mechanism dynamically selects the most informative features, optimizing feature fusion for improved classification. Evaluated on the ADReSSo dataset, MoTAS achieves a leading accuracy of 85.71\%, outperforming existing baselines. Ablation studies further validate the individual contributions of TTS augmentation and MoE in boosting classification performance. These findings highlight the practical value of MoTAS in real-world AD screening scenarios, particularly in data-limited settings.

cs.SD↗

arXiv2025

Reasoning or Not? A Comprehensive Evaluation of Reasoning LLMs for Dialogue Summarization

Dialogue summarization is a challenging task with significant practical value in customer service, meeting analysis, and conversational AI. Although large language models (LLMs) have achieved substantial progress in summarization tasks, the performance of step-by-step reasoning architectures-specifically Long Chain-of-Thought (CoT) implementations such as OpenAI-o1 and DeepSeek-R1-remains unexplored for dialogue scenarios requiring concurrent abstraction and conciseness. In this work, we present the first comprehensive and systematic evaluation of state-of-the-art reasoning LLMs and non-reasoning LLMs across three major paradigms-generic, role-oriented, and query-oriented dialogue summarization. Our study spans diverse languages, domains, and summary lengths, leveraging strong benchmarks (SAMSum, DialogSum, CSDS, and QMSum) and advanced evaluation protocols that include both LLM-based automatic metrics and human-inspired criteria. Contrary to trends in other reasoning-intensive tasks, our findings show that explicit stepwise reasoning does not consistently improve dialogue summarization quality. Instead, reasoning LLMs are often prone to verbosity, factual inconsistencies, and less concise summaries compared to their non-reasoning counterparts. Through scenario-specific analyses and detailed case studies, we further identify when and why explicit reasoning may fail to benefit-or even hinder-summarization in complex dialogue contexts. Our work provides new insights into the limitations of current reasoning LLMs and highlights the need for targeted modeling and evaluation strategies for real-world dialogue summarization.

cs.CL↗

arXiv2025

Wan: Open and Advanced Large-Scale Video Generative Models

This report presents Wan, a comprehensive and open suite of video foundation models designed to push the boundaries of video generation. Built upon the mainstream diffusion transformer paradigm, Wan achieves significant advancements in generative capabilities through a series of innovations, including our novel VAE, scalable pre-training strategies, large-scale data curation, and automated evaluation metrics. These contributions collectively enhance the model's performance and versatility. Specifically, Wan is characterized by four key features: Leading Performance: The 14B model of Wan, trained on a vast dataset comprising billions of images and videos, demonstrates the scaling laws of video generation with respect to both data and model size. It consistently outperforms the existing open-source models as well as state-of-the-art commercial solutions across multiple internal and external benchmarks, demonstrating a clear and significant performance superiority. Comprehensiveness: Wan offers two capable models, i.e., 1.3B and 14B parameters, for efficiency and effectiveness respectively. It also covers multiple downstream applications, including image-to-video, instruction-guided video editing, and personal video generation, encompassing up to eight tasks. Consumer-Grade Efficiency: The 1.3B model demonstrates exceptional resource efficiency, requiring only 8.19 GB VRAM, making it compatible with a wide range of consumer-grade GPUs. Openness: We open-source the entire series of Wan, including source code and all models, with the goal of fostering the growth of the video generation community. This openness seeks to significantly expand the creative possibilities of video production in the industry and provide academia with high-quality video foundation models. All the code and models are available at https://github.com/Wan-Video/Wan2.1.

cs.CV↗

arXiv2025

A Real-time and Hardware Efficient Artfecat-free Spike Sorting Using Deep Spike Detection

Spike sorting is a valuable tool in understanding brain regions. It assigns detected spike waveforms to their origins, helping to research the mechanism of the human brain and the development of implantable brain-machine interfaces (iBMIs). The presence of noise and artefacts will adversely affect the efficacy of spike sorting. This paper proposes a framework for low-cost and real-time implementation of deep spike detection, which consists of two one-dimensional (1-D) convolutional neural network (CNN) model for channel selection and artefact removal. The framework utilizes simulation and hardware layers, and it applies several low-power techniques to optimise the implementation cost of a 1-D CNN model. A compact CNN model with 210 bytes memory size is achieved using structured pruning, network projection and quantization in the simulation layer. The hardware layer also accommodates various techniques including a customized multiply-accumulate (MAC) engine, novel fused layers in the convolution pipeline and proposing flexible resource allocation for a power-efficient and low-delay design. The optimized 1-D CNN significantly decreases both computational complexity and model size, with only a minimal reduction in accuracy. Classification of 1-D CNN on the Cyclone V 5CSEMA5F31C6 FPGA evaluation platform is accomplished in just 16.8 microseconds at a frequency of 2.5 MHz. The FPGA prototype achieves an accuracy rate of 97.14% on a standard dataset and operates with a power consumption of 2.67mW from a supply voltage of 1.1 volts. An accuracy of 95.05% is achieved with a power of 5.6mW when deep spike detection is implemented using two optimized 1-D CNNs on an FPGA board.

eess.SP↗

arXiv2024

LLMCL-GEC: Advancing Grammatical Error Correction with LLM-Driven Curriculum Learning

While large-scale language models (LLMs) have demonstrated remarkable capabilities in specific natural language processing (NLP) tasks, they may still lack proficiency compared to specialized models in certain domains, such as grammatical error correction (GEC). Drawing inspiration from the concept of curriculum learning, we have delved into refining LLMs into proficient GEC experts by devising effective curriculum learning (CL) strategies. In this paper, we introduce a novel approach, termed LLM-based curriculum learning, which capitalizes on the robust semantic comprehension and discriminative prowess inherent in LLMs to gauge the complexity of GEC training data. Unlike traditional curriculum learning techniques, our method closely mirrors human expert-designed curriculums. Leveraging the proposed LLM-based CL method, we sequentially select varying levels of curriculums ranging from easy to hard, and iteratively train and refine using the pretrianed T5 and LLaMA series models. Through rigorous testing and analysis across diverse benchmark assessments in English GEC, including the CoNLL14 test, BEA19 test, and BEA19 development sets, our approach showcases a significant performance boost over baseline models and conventional curriculum learning methodologies.

cs.CL↗

arXiv2024

FOCUS: Forging Originality through Contrastive Use in Self-Plagiarism for Language Models

Pre-trained Language Models (PLMs) have shown impressive results in various Natural Language Generation (NLG) tasks, such as powering chatbots and generating stories. However, an ethical concern arises due to their potential to produce verbatim copies of paragraphs from their training data. This is problematic as PLMs are trained on corpora constructed by human authors. As such, there is a pressing need for research to promote the generation of original content by these models. In this study, we introduce a unique "self-plagiarism" contrastive decoding strategy, aimed at boosting the originality of text produced by PLMs. Our method entails modifying prompts in LLMs to develop an amateur model and a professional model. Specifically, the amateur model is urged to plagiarize using three plagiarism templates we have designed, while the professional model maintains its standard language model status. This strategy employs prompts to stimulate the model's capacity to identify non-original candidate token combinations and subsequently impose penalties. The application of this strategy is integrated prior to the model's final layer, ensuring smooth integration with most existing PLMs (T5, GPT, LLaMA) without necessitating further adjustments. Implementing our strategy, we observe a significant decline in non-original sequences comprised of more than three words in the academic AASC dataset and the story-based ROCStories dataset.

cs.CL↗

arXiv2024

The planar Turán number of $\{K_4,Θ_5\}$

Let $\mathcal{F}$ be a set of graphs. The planar Turán number, $ex_{\mathcal{P}}(n,\mathcal{F})$, is the maximum number of edges in an $n$-vertex planar graph which does not contain any member of $\mathcal{F}$ as a subgraph. In this paper, we give upper bounds of $ex_{\mathcal{P}}(n,\{K_4,Θ_5\})\leqslant25/11(n-2)$. We also give constructions which show the bounds are tight for infinitely many graphs.

math.CO↗

arXiv2023

Multi-Task Instruction Tuning of LLaMa for Specific Scenarios: A Preliminary Study on Writing Assistance

Proprietary Large Language Models (LLMs), such as ChatGPT, have garnered significant attention due to their exceptional capabilities in handling a diverse range of tasks. Recent studies demonstrate that open-sourced smaller foundational models, such as 7B-size LLaMA, can also display remarkable proficiency in tackling diverse tasks when fine-tuned using instruction-driven data. In this work, we investigate a practical problem setting where the primary focus is on one or a few particular tasks rather than general-purpose instruction following, and explore whether LLMs can be beneficial and further improved for such targeted scenarios. We choose the writing-assistant scenario as the testbed, which includes seven writing tasks. We collect training data for these tasks, reframe them in an instruction-following format, and subsequently refine the LLM, specifically LLaMA, via instruction tuning. Experimental results show that fine-tuning LLaMA on writing instruction data significantly improves its ability on writing tasks. We also conduct more experiments and analyses to offer insights for future work on effectively fine-tuning LLaMA for specific scenarios. Finally, we initiate a discussion regarding the necessity of employing LLMs for only one targeted task, taking into account the efforts required for tuning and the resources consumed during deployment.

cs.CL↗

arXiv2023

Extremal graphs for odd-ballooning of paths and stars

The odd-ballooning of a graph $G$, denoted by $G_q$, is the graph obtained from replacing each edge in $G$ by a odd cycle of the same size where the new vertices of the odd cycles are all different. In 2002, Erdös et al. determined the extremal graphs of $k$-fan. In 2016, Hou et al. determined extremal graphs of the odd-ballooning of stars for $q\geqslant 5$. In 2020, Zhu et al. determined extremal graphs of the odd-ballooning of paths for $q\geqslant 3$. In this article, we use progressive induction lemma of Simonovits to determine the extremal graphs of both odd-ballooning of stars and odd-ballooning of paths for $q\geqslant 3$.

math.CO↗

arXiv2023

Turán problems for linear forests and cliques

Given a graph $T$ and a family of graphs $\mathcal{H}$. The generalized Turán number of $\mathcal{H}$ is the maximum number of copies of $T$ in an $\mathcal{H}$-free graph on $n$ vertices, denoted by $ex(n, T, \mathcal{H})$. Let $ex(n, T, \mathcal{H})$ denote the maximum number of copies of $T$ in an $n$-vertex $\mathcal{H}$-free graph. Recently, Alon and Frankl (arXiv2210.15076) determined the exact values of $\rm{ex}(n, \{K_{r+1}, M_{s+1}\})$, where $K_{r+1}$ and $M_{s+1}$ are complete graph on $r + 1$ vertices and matching of size $s + 1$, respectively. Ma and Hou (arXiv2301.05625) gave the generalized version of Alon and Frankl's Theorem, which determine the exact values of $ex(n, K_r, \{K_{k+1}, M_{s+1}\})$. Zhang determined the exact values of $ex(n, K_r, \mathcal{L}_{n, s})$, where $\mathcal{L}_{n, s}$ be the family of all linear forests of order $n$ with $s$ edges. Inspired by the work of Zhang and Ma, in this paper, we determined the exact number of $ex(n, \{K_{r+1}, \mathcal{L}_{n, s}\})$.

math.CO↗

arXiv2023

Is ChatGPT a Highly Fluent Grammatical Error Correction System? A Comprehensive Evaluation

ChatGPT, a large-scale language model based on the advanced GPT-3.5 architecture, has shown remarkable potential in various Natural Language Processing (NLP) tasks. However, there is currently a dearth of comprehensive study exploring its potential in the area of Grammatical Error Correction (GEC). To showcase its capabilities in GEC, we design zero-shot chain-of-thought (CoT) and few-shot CoT settings using in-context learning for ChatGPT. Our evaluation involves assessing ChatGPT's performance on five official test sets in three different languages, along with three document-level GEC test sets in English. Our experimental results and human evaluations demonstrate that ChatGPT has excellent error detection capabilities and can freely correct errors to make the corrected sentences very fluent, possibly due to its over-correction tendencies and not adhering to the principle of minimal edits. Additionally, its performance in non-English and low-resource settings highlights its potential in multilingual GEC tasks. However, further analysis of various types of errors at the document-level has shown that ChatGPT cannot effectively correct agreement, coreference, tense errors across sentences, and cross-sentence boundary errors.

cs.CL↗

arXiv2023

Synergistic Photon Management and Strain-Induced Band Gap Engineering of Two-Dimensional MoS2 Using Semimetal Composite Nanostructures

2D MoS2 attracts increasing attention for its application in flexible electronics and photonic devices. For 2D material optoelectronic devices, light absorption of the molecularly thin 2D absorber would be one of the key limiting factors in device efficiency, and conventional photon management techniques are not necessarily compatible with them. In this paper, we show two semimetal composite nanostructures for synergistic photon management and strain-induced band gap engineering of 2D MoS2: (1) pseudo-periodic Sn nanodots, (2) conductive SnOx (x<1) core-shell nanoneedle structures. Without sophisticated nanolithography, both nanostructures are self-assembled from physical vapor deposition. 2D MoS2 achieves up to >15x enhancement in absorption at λ=650-950 nm under Sn nanodots, and 20-30x at λ=700-900 nm under SnOx (x<1) nanoneedles, both spanning from visible to near infrared regime. Enhanced absorption in MoS2 results from strong near field enhancement and reduced MoS2 band gap due to the tensile strain induced by the Sn nanostructures, as confirmed by Raman and photoluminescence spectroscopy. Especially, we demonstrate that up to 3.5% biaxial tensile strain is introduced to 2D MoS2 using conductive nanoneedle-structured SnOx (x<1), which reduces the band gap by ~0.35 eV to further enhance light absorption at longer wavelengths. To the best of our knowledge, this is the first demonstration of a synergistic triple-functional photon management, stressor, and conductive electrode layer on 2D MoS2. Such synergistic photon management and band gap engineering approach for extended spectral response can be further applied to other 2D materials for future 2D photonic devices.

cond-mat.mes-hall↗

arXiv2022

Some results on the Turán number of $k_1P_{\ell}\cup k_2S_{\ell-1}$

The Turán number of a graph $H$, denoted by $ex(n, H)$, is the maximum number of edges in any graph on $n$ vertices containing no $H$ as a subgraph. Let $P_{\ell}$ denote the path on $\ell$ vertices, $S_{\ell-1}$ denote the star on $\ell$ vertices and $k_1P_{\ell}\cup k_2S_{\ell-1}$ denote the path-star forest with disjoint union of $k_1$ copies of $P_{\ell}$ and $k_2$ copies of $S_{\ell-1}$. In 2013, Lidický et al. first considered the Turán number of $k_1P_4\cup k_2S_3$ for sufficiently large $n$. In 2022, Zhang and Wang raised a conjecture about the Turán number of $k_1P_{2\ell}\cup k_2S_{2\ell-1}$. In this paper, we determine the Turán numbers of $P_{\ell}\cup kS_{\ell-1}$, $k_1P_{2\ell}\cup k_2S_{2\ell-1}$, $2P_5\cup kS_4$ for $n$ appropriately large, which implies the conjecture of Zhang and Wang. The corresponding extremal graphs are also completely characterized.

math.CO↗

arXiv2022

The Turán number of $k_1P_{2l}\cup k_2S_{2l-1}$

The Turán number of a graph $H$, denoted by $ex(n, H)$, is the maximum number of edges in any graph on $n$ vertices containing no $H$ as a subgraph. Let $P_k$ denote the path on $k$ vertices, $S_k$ denote the star on $k+1$ vertices and $k_1P_{2l}\cup k_2S_{2l-1}$ denote the path-star forest with disjoint union of $k_1$ copies of $P_{2l}s$ and $k_2$ copies of $S_{2l-1}s$. In 2019, Lan et al. determined the Turán numbers of $kS_l$ and $k_1P_4\cup k_2S_3$. In 2022, Zhang et al. determined the Turán numbers of $k_1P_6\cup k_2S_5$ and raised a conjecture of the Turán numbers of $k_1P_{2l}\cup k_2S_{2l-1}$, where $k_1\geqslant 1$ and $l\geqslant 2$. In this paper, we study the hypothesis and determine the Turán numbers of $k_1P_{2l}\cup k_2S_{2l-1}$ when $n$ is sufficiently large.

math.CO↗

arXiv2021

Establishing a non-hydrostatic global atmospheric modeling system (iAMAS) at 3-km horizontal resolution with online integrated aerosol feedbacks on the Sunway supercomputer of China

During the era of global warming and highly urbanized development, extreme and high impact weather as well as air pollution incidents influence everyday life and might even cause the incalculable loss of life and property. Although with the vast development of numerical simulation of atmosphere, there still exists substantial forecast biases objectively. To predict extreme weather, severe air pollution, and abrupt climate change accurately, the numerical atmospheric model requires not only to simulate meteorology and atmospheric compositions and their impacts simultaneously involving many sophisticated physical and chemical processes but also at high spatiotemporal resolution. Global atmospheric simulation of meteorology and atmospheric compositions simultaneously at spatial resolutions of a few kilometers remains challenging due to its intensive computational and input/output (I/O) requirement. Through multi-dimension-parallelism structuring, aggressive and finer-grained optimizing, manual vectorizing, and parallelized I/O fragmenting, an integrated Atmospheric Model Across Scales (iAMAS) was established on the new Sunway supercomputer platform to significantly increase the computational efficiency and reduce the I/O cost. The global 3-km atmospheric simulation for meteorology with online integrated aerosol feedbacks with iAMAS was scaled to 39,000,000 processor cores and achieved the speed of 0.82 simulation day per hour (SDPH) with routine I/O, which enables us to perform 5-day global weather forecast at 3-km horizontal resolution with online natural aerosol impacts. The results demonstrate the promising future that the increasing of spatial resolution to a few kilometers with online integrated aerosol impacts may significantly improve the global weather forecast.

physics.ao-ph↗

arXiv2021

Reconstructing Perceptive Images from Brain Activity by Shape-Semantic GAN

Reconstructing seeing images from fMRI recordings is an absorbing research area in neuroscience and provides a potential brain-reading technology. The challenge lies in that visual encoding in brain is highly complex and not fully revealed. Inspired by the theory that visual features are hierarchically represented in cortex, we propose to break the complex visual signals into multi-level components and decode each component separately. Specifically, we decode shape and semantic representations from the lower and higher visual cortex respectively, and merge the shape and semantic information to images by a generative adversarial network (Shape-Semantic GAN). This 'divide and conquer' strategy captures visual information more accurately. Experiments demonstrate that Shape-Semantic GAN improves the reconstruction similarity and image quality, and achieves the state-of-the-art image reconstruction performance.

cs.NE↗

arXiv2020

Integrated Design of Unmanned Aerial Mobility Network: A Data-Driven Risk-Averse Approach

The real challenge in drone-logistics is to develop an economically-feasible Unmanned Aerial Mobility Network (UAMN). In this paper, we propose an integrated airport location (strategic decision) and routes planning (operational decision) optimization framework to minimize the total cost of the network, while guaranteeing flow constraints, capacity constraints, and electricity constraints. To facility expensive long-term infrastructure planning facing demand uncertainty, we develop a data-driven risk-averse two-stage stochastic optimization model based on the Wasserstein distance. We develop a reformulation technique which simplifies the worst-case expectation term in the original model, and obtain a fractable Min-Max solution procedure correspondingly. Using Lagrange multipliers, we successfully decompose decision variables and reduce the complexity of computation. To provide managerial insights, we design specific numerical examples. For example, we find that the optimal network configuration is affected by the "pooling effects" in channel capacities. A nice feature of our DRO framework is that the optimal network design is relatively robust under demand uncertainty. Interestingly, a candidate node without historical demand records can be chosen to locate an airport. We demonstrate the application of our model for a real medical resources transportation problem with our industry partner, collecting donated blood to a blood bank in Hangzhou, China.

math.OC↗

arXiv2019

Silicon Nitride Stress Liner Impacts on the Electrical Characteristics of AlGaN/GaN HEMTs

Due to the piezoelectric nature of GaN, the 2DEG in AlGaN/GaN HEMT could be engineered by strain. In this work, SiNx deposited using dual-frequency PECVD was used as a stressor. The output performance of the devices was dominated by the surface passivation instead of the stress effect. However, the threshold voltage was increased by the induced stress, supporting strain engineering as an effective approach to pursue the normally-off operation of AlGaN/GaN HEMTs.

physics.app-ph↗