arXiv subjects

Shumin Zhang

Publications and source records attributed to Shumin Zhang.

7 recordsLinked to original sources

arXiv2026

Fronto-parietal and fronto-temporal EEG coherence as predictive neuromarkers of transcutaneous auricular vagus nerve stimulation response in treatment-resistant schizophrenia: A machine learning study

Response variability limits the clinical utility of transcutaneous auricular vagus nerve stimulation (taVNS) for negative symptoms in treatment-resistant schizophrenia (TRS). This study aimed to develop an electroencephalography (EEG)-based machine learning (ML) model to predict individual response and explore associated neurophysiological mechanisms. We used ML to develop and validate predictive models based on pre-treatment EEG data features (power, coherence, and dynamic functional connectivity) from 50 TRS patients enrolled in the taVNS trial, within a nested cross-validation framework. Participants received 20 sessions of active or sham taVNS (n = 25 each) over two weeks, followed by a two-week follow-up. The prediction target was the percentage change in the positive and negative syndrome scale-factor score for negative symptoms (PANSS-FSNS) from baseline to post-treatment, with further evaluation of model specificity and neurophysiological relevance.The optimal model accurately predicted taVNS response in the active group, with predicted PANSS-FSNS changes strongly correlated with observed changes (r = 0.87, p < .001); permutation testing confirmed performance above chance (p < .001). Nine consistently retained features were identified, predominantly fronto-parietal and fronto-temporal coherence features. Negligible predictive performance in the sham group and failure to predict positive symptom change support the predictive specificity of this oscillatory signature for taVNS-related negative symptom improvement. Two coherence features within fronto-parietal-temporal networks showed post-taVNS changes significantly associated with symptom improvement, suggesting dual roles as predictors and potential therapeutic targets. EEG oscillatory neuromarkers enable accurate prediction of individual taVNS response in TRS, supporting mechanism-informed precision neuromodulation strategies.

cs.LG↗

arXiv2026

The Side Effects of Being Smart: Safety Risks in MLLMs' Multi-Image Reasoning

As Multimodal Large Language Models (MLLMs) acquire stronger reasoning capabilities to handle complex, multi-image instructions, this advancement may pose new safety risks. We study this problem by introducing MIR-SafetyBench, the first benchmark focused on multi-image reasoning safety, which consists of 2,676 instances across a taxonomy of 9 multi-image relations. Our extensive evaluations on 19 MLLMs reveal a troubling trend: models with more advanced multi-image reasoning can be more vulnerable on MIR-SafetyBench. Beyond attack success rates, we find that many responses labeled as safe are superficial, often driven by misunderstanding or evasive, non-committal replies. We further observe that unsafe generations exhibit lower attention entropy than safe ones on average. This internal signature suggests a possible risk that models may over-focus on task solving while neglecting safety constraints. Our code and data are available at https://github.com/thu-coai/MIR-SafetyBench.

cs.CV↗

arXiv2025

Reliability evaluation of Cayley graph generated by unicyclic graphs based on cyclic fault pattern

Graph connectivity serves as a fundamental metric for evaluating the reliability and fault tolerance of interconnection networks. To more precisely characterize network robustness, the concept of cyclic connectivity has been introduced, requiring that there are at least two components containing cycles after removing the vertex set. This property ensures the preservation of essential cyclic communication structures under faulty conditions. Cayley graphs exhibit several ideal properties for interconnection networks, which permits identical routing protocols at all vertices, facilitates recursive constructions, and ensures operational robustness. In this paper, we investigate the cyclic connectivity of Cayley graphs generated by unicyclic triangle free graphs. Given an symmetric group $Sym(n)$ on $\left\{ 1,2,\dots,n\right\}$ and a set $\mathcal{T}$ of transpositions of $Sym(n)$. Let $G(\mathcal{T})$ be the graph on vertex set $\left\{ 1,2,\dots,n\right\}$ and edge set $\left\{ij\colon(ij)\in \mathcal{T}\right\}$. If $G(\mathcal{T})$ is a unicyclic triangle free graphs, then denoted the Cayley graph Cay$(Sym(n),\mathcal{T})$ by $UG_{n}$. As a result, we determine the exact value of cyclic connectivity of $UG_{n}$ as $\kappa_{c}(UG_{n})=4n-8$ for $n\ge 4 $.

math.CO↗

arXiv2025

Spectral Sufficient Conditions for Graph Factors

The $\{K_{1,1}, K_{1,2},C_m: m\geq3\}$-factor of a graph is a spanning subgraph whose each component is an element of $\{K_{1,1}, K_{1,2},C_m: m\geq3\}$. In this paper, through the graph spectral methods, we establish the lower bound of the signless Laplacian spectral radius and the upper bound of the distance spectral radius to determine whether a graph admits a $\{K_2\}$-factor. We get a lower bound on the size (resp. the spectral radius) of $G$ to guarantee that $G$ contains a $\{K_{1,1}, K_{1,2},C_m: m\geq3\}$-factor. Then we determine an upper bound on the distance spectral radius of $G$ to ensure that $G$ has a $\{K_{1,1}, K_{1,2},C_m: m\geq3\}$-factor. Furthermore, by constructing extremal graphs, we show that the above all bounds are best possible.

math.CO↗

arXiv2025

Partial domination of middle graphs

For any graph $G=(V,E)$, a subset $S\subseteq V$ is called {\it an isolating set} of $G$ if $V\setminus N_G[S]$ is an independent set of $G$, where $N_G[S]=S\cup N_G(S)$, and {\it the isolation number} of $G$, denoted by $\iota(G)$, is the size of a smallest isolating set of $G$. In this article, we show that the isolation number of the middle graph of $G$ is equal to the size of a smallest maximal matching of $G$.

math.CO↗

arXiv2020

The g-extra connectivity of the Mycielskian

The $g$-extra connectivity is an important parameter to measure the ability of tolerance and reliability of interconnection networks. Given a connected graph $G=(V,E)$ and a non-negative integer $g$, a subset $S\subseteq V$ is called a $g$-extra cut of $G$ if $G-S$ is disconnected and every component of $G-S$ has at least $g+1$ vertices. The cardinality of the minimum $g$-extra cut is defined as the $g$-extra connectivity of $G$, denoted by $\kappa_g(G)$. In a search for triangle-free graphs with arbitrarily large chromatic numbers, Mycielski developed a graph transformation that transforms a graph $G$ into a new graph $\mu(G)$, which is called the Mycielskian of $G$. This paper investigates the relationship of the g-extra connectivity of the Mycielskian $\mu(G)$ and the graph $G$, moreover, show that $\kappa_{2g+1}(\mu(G))=2\kappa_{g}(G)+1$ for $g\geq 1$ and $\kappa_{g}(G)\leq min\{g+1, \lfloor\frac{n}{2}\rfloor\}$.

math.CO↗

arXiv2019

Controllable pulse patterns in fiber lasers

An all-optical pulse power editing (PPE) technique is reported. Using the PPE technique, pulses with different peak powers are output and directed to the positive or reverse saturable absorption (SA) range of the saturable absorber. Further, under the combined action of the PPE technique and SA, four pulse patterns including dissipative soliton (DS), DS molecules, soliton compounds composed of DS and noise-like pulse (NLP), and pure NLP are controllably generated in fiber lasers. The results are conducive for developing advanced DS lasers and can further clarify the onset of pulse dynamic patterns.

physics.optics↗