搜索结果：BioImpacts : BI

Vision Language Model (VLM) typically assume complete modality input during inference. However, their effectiveness drops sharply when certain modalities are unavailable or incomplete. Current research on missing modality primarily faces two dilemmas: Prompt-based methods struggle to restore missing yet indispensable features and degrade the generalizability of VLM. Imputation-based approaches, lacking effective guidance, are prone to generating semantically irrelevant noise. Restoring precise semantics while sustaining VLM's generalization remains challenging. Therefore, we propose a general missing modality restoration strategy in this paper. We introduce an enhanced diffusion model as a pluggable mid-stage training module to effectively restore missing features. Our strategy introduces two key innovations: (I) Dynamic Modality Gating, which adaptively leverages conditional features to guide the generation of semantically consistent features; (II) Cross-Modal Mutual Learning mechanism, which bridges the semantic spaces of the dual models to achieve bi-directional alignment. Notably, our strategy maintains the original integrity of the pre-trained VLM, requiring no fine-tuning of

This paper proposes a hybrid mono- and bi-static sensing framework, by leveraging the base station (BS) and user equipment (UE) cooperation in integrated sensing and communication (ISAC) systems. This scheme is built on 3GPP-supported sensing modes, and it does not incur any extra spectrum cost or inter-cell coordination. To reveal the fundamental performance limit of the proposed hybrid sensing mode, we derive closed-form Cramér-Rao lower bound (CRLB) for sensing target localization and velocity estimation, as functions of target and UE positions. The results reveal that significant performance gains can be achieved over the purely mono- or bi-static sensing, especially when the BS-target-UE form a favorable geometry, which is close to a right triangle. The analytical results are validated by simulations using effective parameter estimation algorithm and weighted mean square error (MSE) fusion method. Based on the derived sensing bound, we further analyze the sensing coverage by varying the UE positions, which shows that sensing coverage first improves then degrades as the BS-UE separation increases. Furthermore, the sensing accuracy for a potential target with best UE selection i

We report a series of high-pressure electrical transport, magnetic susceptibility, and x-ray diffraction measurements on single crystals of the weak topological insulator Bi2TeI and the topological metal Bi3TeI. Room temperature x-ray diffraction measurements show that both materials go through a series of pressure-induced structural transitions and eventually adopt a disordered bcc alloy structure at high pressure. A re-analysis of the published data on BiTeI indicates that this material also adopts a disordered bcc structure at high pressure, in contrast to the previously suggested P4/nmm structure. We find that Bi2TeI and Bi3TeI become superconducting at 13 GPa and 11.5 GPa, respectively. The superconducting critical temperature Tc of the bcc phase reaches maximum values of 7 K and 7.5 K for Bi2TeI and Bi3TeI, respectively and dTc/dP < 0 in both cases. The results indicate that disordered alloy bcc superconducting phases appear to be a universal feature of both the Bi-Te and Bi-Te-I systems at high pressure.

The investigation of crystallographic, electronic, and magnetic characteristics, especially the mixed valences of Eu$^{2+}$ and Eu$^{3+}$ under pressure of a novel europium-based bismuth selenide compound, Eu$_4$Bi$_6$Se$_{13}$, presented. This new compound adopts a monoclinic crystal structure classified under the P$2_1$/m space group (#11). It exhibits distinctive structural features, including substantial Eu-Se coordination numbers, Bi-Se ladders, and linear chains of Eu atoms that propagate along the b-axis. Electronic resistivity assessments indicate that Eu$_{4}$Bi$_{6}$Se$_{13}$ exhibits weak metallic behaviors. Magnetic characterization reveals uniaxial magnetic anisotropy, with a notable spin transition at approximately 1.2 T when the magnetic field is oriented along the b-axis. This behavior, coupled with the specific Eu-Eu interatomic distances and the magnetic saturation observed at low fields, supports the identification of metamagnetic properties attributable to the flipping of europium spins. The Curie-Weiss analysis of the magnetic susceptibility measured both perpendicular and parallel to the b-axis and high-pressure partial fluorescence yield (PFY) results detecte

Dwarf spheroidal galaxies (dSphs) are excellent targets for indirect dark matter (DM) searches using gamma-ray telescopes because they are thought to have high DM content and a low astrophysical background. The sensitivity of these searches is improved by combining the observations of dSphs made by different gamma-ray telescopes. We present the results of a combined search by the most sensitive currently operating gamma-ray telescopes, namely: the satellite-borne Fermi-LAT telescope; the ground-based imaging atmospheric Cherenkov telescope arrays H.E.S.S., MAGIC, and VERITAS; and the HAWC water Cherenkov detector. Individual datasets were analyzed using a common statistical approach. Results were subsequently combined via a global joint likelihood analysis. We obtain constraints on the velocity-weighted cross section $\langle σ\mathit{v} \rangle$ for DM self-annihilation as a function of the DM particle mass. This five-instrument combination allows the derivation of up to 2-3 times more constraining upper limits on $\langle σ\mathit{v} \rangle$ than the individual results over a wide mass range spanning from 5 GeV to 100 TeV. Depending on the DM content modeling, the 95% confidence

Remote photoplethysmography (rPPG) is gaining prominence for its non-invasive approach to monitoring physiological signals using only cameras. Despite its promise, the adaptability of rPPG models to new, unseen domains is hindered due to the environmental sensitivity of physiological signals. To address this, we pioneer the Test-Time Adaptation (TTA) in rPPG, enabling the adaptation of pre-trained models to the target domain during inference, sidestepping the need for annotations or source data due to privacy considerations. Particularly, utilizing only the user's face video stream as the accessible target domain data, the rPPG model is adjusted by tuning on each single instance it encounters. However, 1) TTA algorithms are designed predominantly for classification tasks, ill-suited in regression tasks such as rPPG due to inadequate supervision. 2) Tuning pre-trained models in a single-instance manner introduces variability and instability, posing challenges to effectively filtering domain-relevant from domain-irrelevant features while simultaneously preserving the learned information. To overcome these challenges, we present Bi-TTA, a novel expert knowledge-based Bidirectional Tes

In this paper, we extend the notion of microstate free entropy to the bi-free setting. In particular, using the bi-free analogue of random matrices, microstate bi-free entropy is defined. Properties essential to an entropy theory are developed, such as the behaviour of the entropy when transformations on the left variables or on the right variables are performed. In addition, the microstate bi-free entropy is demonstrated to be additive over bi-free collections provided additional regularity assumptions are included and is computed for all bi-free central limit distributions. Moreover, an orbital version of bi-free entropy is examined which provides a tighter upper bound for the subadditivity of microstate bi-free entropy and provides an alternate characterization of bi-freeness in certain settings.

Open-vocabulary change detection aims to identify semantic changes in bi-temporal remote sensing images without predefined categories. Recent methods combine foundation models such as SAM, DINO and CLIP, but typically process each timestamp independently or interact only at the final comparison stage. Such paradigms suffer from insufficient temporal coupling during semantic reasoning, which limits their ability to distinguish genuine semantic changes from non-semantic appearance discrepancies. In addition, patch-dominant inference on high-resolution images often weakens global semantic continuity and produces fragmented change regions. To address these issues, we propose MemOVCD, a training-free open-vocabulary change detection framework based on cross-temporal memory reasoning and global-local adaptive rectification. Specifically, we reformulate bi-temporal change detection as a two-frame tracking problem and introduce weighted bidirectional propagation to aggregate semantic evidence from both temporal directions. To stabilize memory propagation across large temporal gaps, we construct histogram-aligned transition frames to smooth abrupt appearance changes. Moreover, a global-loca

In this work, we study the 1-bit convolutional neural networks (CNNs), of which both the weights and activations are binary. While being efficient, the classification accuracy of the current 1-bit CNNs is much worse compared to their counterpart real-valued CNN models on the large-scale dataset, like ImageNet. To minimize the performance gap between the 1-bit and real-valued CNN models, we propose a novel model, dubbed Bi-Real net, which connects the real activations (after the 1-bit convolution and/or BatchNorm layer, before the sign function) to activations of the consecutive block, through an identity shortcut. Consequently, compared to the standard 1-bit CNN, the representational capability of the Bi-Real net is significantly enhanced and the additional cost on computation is negligible. Moreover, we develop a specific training algorithm including three technical novelties for 1- bit CNNs. Firstly, we derive a tight approximation to the derivative of the non-differentiable sign function with respect to activation. Secondly, we propose a magnitude-aware gradient with respect to the weight for updating the weight parameters. Thirdly, we pre-train the real-valued CNN model with a

In this paper, we study 1-bit convolutional neural networks (CNNs), of which both the weights and activations are binary. While efficient, the lacking of representational capability and the training difficulty impede 1-bit CNNs from performing as well as real-valued networks. We propose Bi-Real net with a novel training algorithm to tackle these two challenges. To enhance the representational capability, we propagate the real-valued activations generated by each 1-bit convolution via a parameter-free shortcut. To address the training difficulty, we propose a training algorithm using a tighter approximation to the derivative of the sign function, a magnitude-aware gradient for weight updating, a better initialization method, and a two-step scheme for training a deep network. Experiments on ImageNet show that an 18-layer Bi-Real net with the proposed training algorithm achieves 56.4% top-1 classification accuracy, which is 10% higher than the state-of-the-arts (e.g., XNOR-Net) with greater memory saving and lower computational cost. Bi-Real net is also the first to scale up 1-bit CNNs to an ultra-deep network with 152 layers, and achieves 64.5% top-1 accuracy on ImageNet. A 50-layer

One-Shot Federated Learning (OSFL), a special decentralized machine learning paradigm, has recently gained significant attention. OSFL requires only a single round of client data or model upload, which reduces communication costs and mitigates privacy threats compared to traditional FL. Despite these promising prospects, existing methods face challenges due to client data heterogeneity and limited data quantity when applied to real-world OSFL systems. Recently, Latent Diffusion Models (LDM) have shown remarkable advancements in synthesizing high-quality images through pretraining on large-scale datasets, thereby presenting a potential solution to overcome these issues. However, directly applying pretrained LDM to heterogeneous OSFL results in significant distribution shifts in synthetic data, leading to performance degradation in classification models trained on such data. This issue is particularly pronounced in rare domains, such as medical imaging, which are underrepresented in LDM's pretraining data. To address this challenge, we propose Federated Bi-Level Personalization (FedBiP), which personalizes the pretrained LDM at both instance-level and concept-level. Hereby, FedBiP sy

We report the analysis of microlensing event OGLE-2017-BLG-1038, observed by the Optical Gravitational Lensing Experiment, Korean Microlensing Telescope Network, and Spitzer telescopes. The event is caused by a giant source star in the Galactic Bulge passing over a large resonant binary lens caustic. The availability of space-based data allows the full set of physical parameters to be calculated. However, there exists an eightfold degeneracy in the parallax measurement. The four best solutions correspond to very-low-mass binaries near ($M_1 = 170^{+40}_{-50} M_J$ and $M_2 = 110^{+20}_{-30} M_J$), or well below ($M_1 = 22.5^{+0.7}_{-0.4} M_J$ and $M_2 = 13.3^{+0.4}_{-0.3} M_J$) the boundary between stars and brown dwarfs. A conventional analysis, with scaled uncertainties for Spitzer data, implies a very-low-mass brown dwarf binary lens at a distance of 2 kpc. Compensating for systematic Spitzer errors using a Gaussian process model suggests that a higher mass M-dwarf binary at 6 kpc is equally likely. A Bayesian comparison based on a galactic model favors the larger-mass solutions. We demonstrate how this degeneracy can be resolved within the next ten years through infrared adaptiv

In contrast to the robust mutual interpretability phenomenon in set theory, Ali Enayat proved that bi-interpretation is absent: distinct theories extending ZF are never bi-interpretable and models of ZF are bi-interpretable only when they are isomorphic. Nevertheless, for natural weaker set theories, we prove, including Zermelo-Fraenkel set theory $\text{ZFC}^-$ without power set and Zermelo set theory Z, there are nontrivial instances of bi-interpretation. Specifically, there are well-founded models of $\text{ZFC}^-$ that are bi-interpretable, but not isomorphic---even $\langle H_{ω_1},\in\rangle$ and $\langle H_{ω_2},\in\rangle$ can be bi-interpretable---and there are distinct bi-interpretable theories extending $\text{ZFC}^-$. Similarly, using a construction of Mathias, we prove that every model of ZF is bi-interpretable with a model of Zermelo set theory in which the replacement axiom fails.

For the nonlinear Dirac equation with scalar self-interaction (the Soler model) in three spatial dimensions, we consider the linearization at solitary wave solutions and find the invariant spaces which correspond to different spherical harmonics, thus achieving the radial reduction of the spectral stability analysis. We apply the same technique to the bi-frequency solitary waves (which are generically present in the Soler model) and show that they can also possess linear stability properties similar to those of one-frequency solitary waves.

We present an overview of the mathematical structure of geminal theory within the seniority formalism and bi-variational principle. Named after the constellation, geminal wavefunctions provide the mean-field like representation of paired-electron wavefunctions in quantum chemistry, tying in with the Lewis picture of chemical bonding via electron pairs. Unfortunately, despite its mean-field product wave function description, the computational cost of computing geminal wavefunctions is dominated by the permanent overlaps with Slater determinant reference states. We review recent approaches to reduce the factorial scaling of the permanent, and present the bi-variational principle as a consistent framework for the projected Schrödinger Equation and the computation of reduced density matrices.

We study the properties of the analogue of R-diagonal operators in the setting of bi-free probability. Products of bi-R-diagonal pairs of operators that are $*$-bi-free are studied and powers of such pairs are found to also be bi-R-diagonal. It is moreover shown that the joint $*$-distribution of a bi-R-diagonal pair of operators remains invariant under the multiplication by a $*$-bi-free bi-Haar unitary pair and equivalent characterizations of bi-R-diagonal pairs are developed.

Boolean function bi-decomposition is ubiquitous in logic synthesis. It entails the decomposition of a Boolean function using two-input simple logic gates. Existing solutions for bi-decomposition are often based on BDDs and, more recently, on Boolean Satisfiability. In addition, the partition of the input set of variables is either assumed, or heuristic solutions are considered for finding good partitions. In contrast to earlier work, this paper proposes the use of Quantified Boolean Formulas (QBF) for computing bi- decompositions. These bi-decompositions are optimal in terms of the achieved disjointness and balancedness of the input set of variables. Experimental results, obtained on representative benchmarks, demonstrate clear improvements in the quality of computed decompositions, but also the practical feasibility of QBF-based bi-decomposition.

We develop a theoretical model to calculate the quantum efficiency (QE) of photoelectron emission from materials with Rashba spin-orbit coupling (RSOC) effect. In the low temperature limit, an analytical scaling between QE and the RSOC strength is obtained as QE $\propto (\hbarω-W)^2+2E_R(\hbar ω-W) -E_R^2/3$, where $\hbarω$, $W$ and $E_R$ are the incident photon energy, work function and the RSOC parameter respectively. Intriguingly, the RSOC effect substantially improves the QE for strong RSOC materials. For example, the QE of Bi$_2$Se$_3$ and Bi/Si(111) increases, by 149\% and 122\%, respectively due to the presence of strong RSOC. By fitting to the photoelectron emission characteristics, the analytical scaling law can be employed to extract the RSOC strength, thus offering a useful tool to characterize the RSOC effect in materials. Importantly, when the traditional Fowler-Dubridge model is used, the extracted results may substantially deviate from the actual values by $\sim90\%$, thus highlighting the importance of employing our model to analyse the photoelectron emission especially for materials with strong RSOC. These findings provide a theoretical foundation for the design o

Bi-level optimization, especially the gradient-based category, has been widely used in the deep learning community including hyperparameter optimization and meta-knowledge extraction. Bi-level optimization embeds one problem within another and the gradient-based category solves the outer-level task by computing the hypergradient, which is much more efficient than classical methods such as the evolutionary algorithm. In this survey, we first give a formal definition of the gradient-based bi-level optimization. Next, we delineate criteria to determine if a research problem is apt for bi-level optimization and provide a practical guide on structuring such problems into a bi-level optimization framework, a feature particularly beneficial for those new to this domain. More specifically, there are two formulations: the single-task formulation to optimize hyperparameters such as regularization parameters and the distilled data, and the multi-task formulation to extract meta-knowledge such as the model initialization. With a bi-level formulation, we then discuss four bi-level optimization solvers to update the outer variable including explicit gradient update, proxy update, implicit functi