Search

The rapid advancement of Large Vision-Language Models (LVLMs) has enhanced capabilities offering potential applications from content creation to productivity enhancement. Despite their innovative potential, LVLMs exhibit vulnerabilities, especially in generating potentially toxic or unsafe responses. Malicious actors can exploit these vulnerabilities to propagate toxic content in an automated (or semi-) manner, leveraging the susceptibility of LVLMs to deception via strategically crafted prompts without fine-tuning or compute-intensive procedures. Despite the red-teaming efforts and inherent potential risks associated with the LVLMs, exploring vulnerabilities of LVLMs remains nascent and yet to be fully addressed in a systematic manner. This study systematically examines the vulnerabilities of open-source LVLMs, including LLaVA, InstructBLIP, Fuyu, and Qwen, using adversarial prompt strategies that simulate real-world social manipulation tactics informed by social theories. Our findings show that (i) toxicity and insulting are the most prevalent behaviors, with the mean rates of 16.13% and 9.75%, respectively; (ii) Qwen-VL-Chat, LLaVA-v1.6-Vicuna-7b, and InstructBLIP-Vicuna-7b are the most vulnerable models, exhibiting toxic response rates of 21.50%, 18.30% and 17.90%, and insulting responses of 13.40%, 11.70% and 10.10%, respectively; (iii) prompting strategies incorporating dark humor and multimodal toxic prompt completion significantly elevated these vulnerabilities. Despite being fine-tuned for safety, these models still generate content with varying degrees of toxicity when prompted with adversarial inputs, highlighting the urgent need for enhanced safety mechanisms and robust guardrails in LVLM development.

There is a growing interest in codebook-based beam-steering for millimeter-wave (mmWave) systems due to its potential for low complexity and rapid beam search. A key focus of recent research has been the design of codebooks that strike a trade-off between achievable gain and codebook size, which directly impacts beam search time. Statistical approaches have shown promise by leveraging the likelihood that certain beam directions (equivalently, sets of phase-shifter configurations) are more probable than others. Such approaches are shown to be valid for static, non-rotating transmission stations such as base stations. However, for the case of user terminals that are constantly changing orientation, the possible phase-shifter configurations become equally probable, rendering statistical methods less relevant. On the other hand, user terminals come with a large number of possible steering vector configurations, which can span up to six orders of magnitude. Therefore, efficient solutions to reduce the codebook size (set of possible steering vectors) without compromising array gain are needed. We address this challenge by proposing a novel and practical codebook refinement technique, aiming to reduce the codebook size while maintaining array gain within $\gamma$ dB of the maximum achievable gain at any random orientation of the user terminal. We project that a steering vector at a given angle could effectively cover adjacent angles with a small gain loss compared to the maximum achievable gain. We demonstrate experimentally that it is possible to reduce the codebook size from $1024^{16}$ to just a few configurations (e.g., less than ten), covering all angles while maintaining the gain within $\gamma=3$ dB of the maximum achievable gain., Comment: 6 pages, 3 figures, conference

In the present paper, we deal with Bernstein-Chlodowsky type operators for approximating functions on the domain. We first present Bernstein-Chlodowsky type operators in two variables and then we discuss some examples of these operators under a domain transformation. Finally, we give bivarite shifted mth Bernstein-Chlodowsky-Stancu operators and we present some figures for approximation properties of our operator.

The rapid spread of misinformation on social media, especially during crises, challenges public decision-making. To address this, we propose HierTKG, a framework combining Temporal Graph Networks (TGN) and hierarchical pooling (DiffPool) to model rumor dynamics across temporal and structural scales. HierTKG captures key propagation phases, enabling improved temporal link prediction and actionable insights for misinformation control. Experiments demonstrate its effectiveness, achieving an MRR of 0.9845 on ICEWS14 and 0.9312 on WikiData, with competitive performance on noisy datasets like PHEME (MRR: 0.8802). By modeling structured event sequences and dynamic social interactions, HierTKG adapts to diverse propagation patterns, offering a scalable and robust solution for real-time analysis and prediction of rumor spread, aiding proactive intervention strategies., Comment: Preprint

The purpose of this paper is to obtain Fourier transforms of multivariate orthogonal polynomials on the cone such as Laguerre polynomials on the cone and Jacobi polynomials on the cone and to define two new families of multivariate orthogonal functions by using Parseval's identity. Also, the obtained results are expressed in terms of the continuous Hahn polynomials., Comment: 22 pages. arXiv admin note: text overlap with arXiv:2206.13893

The first spectroscopy of $^{52}$K was investigated via in-beam $\gamma$-ray spectroscopy at the RIKEN Radioactive Isotope Beam Factory after one-proton and one-neutron knockout from $^{53}$Ca and $^{53}$K beams impinging on a 15-cm liquid hydrogen target at $\approx$ 230~MeV/nucleon. The energy level scheme of $^{52}$K was built using single $\gamma$ and $\gamma$-$\gamma$ coincidence spectra. The spins and parities of the excited states were established based on momentum distributions of the fragment after the knockout reaction and based on exclusive cross sections. The results were compared to state-of-the-art shell model calculations with the SDPF-Umod interaction and ab initio IMSRG calculations with chiral effective field theory nucleon-nucleon and three-nucleon forces.

In invasive breast cancer, cell clusters of varying sizes and shapes are embedded in the fibrous extracellular matrix (ECM). Although the prevailing view attributes this structure to increasing disorder resulting from loss of function and dedifferentiation, our findings reveal that it arises through a process of active self-organization driven by cancer cell motility. Simulations and histological analyses of tumours from over 2,000 breast cancer patients reveal that motile, aligned cancer cells within clusters move as active nematic aggregates through the surrounding highly aligned ECM fibres, which form a confining, passive nematic phase. Cellular motion leads to cluster splitting and coalescence. The degree of cluster activity, combined with heterogeneity in cell motility, is reflected in specific scaling behaviours for cluster shape, size distribution, and the distance between cluster boundaries and nematic defects in ECM alignment. Increased activity estimates correlate with tumour progression and are associated with a poorer prognosis for patients.

Accurate classification of Intraductal Papillary Mucinous Neoplasms (IPMN) is essential for identifying high-risk cases that require timely intervention. In this study, we develop a federated learning framework for multi-center IPMN classification utilizing a comprehensive pancreas MRI dataset. This dataset includes 653 T1-weighted and 656 T2-weighted MRI images, accompanied by corresponding IPMN risk scores from 7 leading medical institutions, making it the largest and most diverse dataset for IPMN classification to date. We assess the performance of DenseNet-121 in both centralized and federated settings for training on distributed data. Our results demonstrate that the federated learning approach achieves high classification accuracy comparable to centralized learning while ensuring data privacy across institutions. This work marks a significant advancement in collaborative IPMN classification, facilitating secure and high-accuracy model training across multiple centers.

Federated learning (FL) enables collaborative model training across institutions without sharing sensitive data, making it an attractive solution for medical imaging tasks. However, traditional FL methods, such as Federated Averaging (FedAvg), face difficulties in generalizing across domains due to variations in imaging protocols and patient demographics across institutions. This challenge is particularly evident in pancreas MRI segmentation, where anatomical variability and imaging artifacts significantly impact performance. In this paper, we conduct a comprehensive evaluation of FL algorithms for pancreas MRI segmentation and introduce a novel approach that incorporates adaptive aggregation weights. By dynamically adjusting the contribution of each client during model aggregation, our method accounts for domain-specific differences and improves generalization across heterogeneous datasets. Experimental results demonstrate that our approach enhances segmentation accuracy and reduces the impact of domain shift compared to conventional FL methods while maintaining privacy-preserving capabilities. Significant performance improvements are observed across multiple hospitals (centers).

Leading open-source large language models (LLMs) such as Llama-3.1-Instruct-405B are extremely capable at generating text, answering questions, and solving a variety of natural language understanding tasks. However, they incur higher inference cost and latency compared to smaller LLMs. Knowledge distillation provides a way to use outputs from these large, capable teacher models to train smaller student models which can be used for inference at lower cost and latency, while retaining comparable accuracy. We investigate the efficacy of distillation using the Llama-3.1-405B-Instruct teacher and the smaller Llama-3.1-8B-Instruct and Llama-3.1-70B-Instruct student models. Contributions of this work include (a) We evaluate the generalizability of distillation with the above Llama-3.1 teacher-student pairs across different tasks and datasets (b) We show that using synthetic data during distillation significantly improves the accuracy of 8B and 70B models, and when used with reasoning chains, even matches or surpasses the zero-shot accuracy of 405B model on some datasets (c) We empirically show that distillation enables 8B and 70B models to internalize 405B's reasoning ability by using only standard fine-tuning (without customizing any loss function). This allows cost and latency-efficient student model inference. (d) We show pitfalls in evaluation of distillation, and present task-specific evaluation, including both human and LLM-grading, and ground-truth based traditional accuracy benchmarks. This methodical study brings out the fundamental importance of synthetic data quality in knowledge distillation, and of combining multiple, task-specific ways of accuracy and quality evaluation in assessing the effectiveness of distillation., Comment: 25 pages, 4 figures

This paper proposes a novel neural network model capable of discovering high-level skill representations from unlabeled demonstration data. We also propose a bi-level planning pipeline that utilizes our model using a gradient-based planning approach. While extracting high-level representations, our model also preserves the low-level information, which can be used for low-level action planning. In the experiments, we tested the skill discovery performance of our model under different conditions, tested whether Multi-Modal LLMs can be utilized to label the learned high-level skill representations, and finally tested the high-level and low-level planning performance of our pipeline., Comment: 12 pages, 6 figures, Submitted to Conference on Robot Learning LEAP Workshop 2024

Liver cirrhosis, the end stage of chronic liver disease, is characterized by extensive bridging fibrosis and nodular regeneration, leading to an increased risk of liver failure, complications of portal hypertension, malignancy and death. Early diagnosis and management of end-stage cirrhosis are significant clinical challenges. Magnetic resonance imaging (MRI) is a widely available, non-invasive imaging technique for cirrhosis assessment. However, the stage of liver fibrosis cannot be easily differentiated. Moreover, the fibrotic liver tissue (cirrhotic liver) causes significant change in liver enhancement, morphology and signal characteristics, which poses substantial challenges for the development of computer-aided diagnostic applications. Deep learning (DL) offers a promising solution for automatically segmenting and recognizing cirrhotic livers in MRI scans, potentially enabling fibrosis stage classification. However, the lack of datasets specifically focused on cirrhotic livers has hindered progress. CirrMRI600+ addresses this critical gap. This extensive dataset, the first of its kind, comprises 628 high-resolution abdominal MRI scans (310 T1-weighted and 318 T2-weighted, totaling nearly 40,000 slices) with annotated segmentation labels for cirrhotic livers. Unlike previous datasets, CirrMRI600+ specifically focuses on cirrhotic livers, capturing the complexities of this disease state. The link to the dataset is made publicly available at: https://osf.io/cuk24/. We also share 11 baseline deep learning segmentation methods used in our rigorous benchmarking experiments: https://github.com/NUBagciLab/CirrMRI600Plus.

This paper considers the minimization of a continuously differentiable function over a cardinality constraint. We focus on smooth and relatively smooth functions. These smoothness criteria result in new descent lemmas. Based on the new descent lemmas, novel optimality conditions and algorithms are developed, which extend the previously proposed hard-thresholding algorithms. We give a theoretical analysis of these algorithms and extend previous results on properties of iterative hard thresholding-like algorithms. In particular, we focus on the weighted $\ell_2$ norm, which requires efficient solution of convex subproblems. We apply our algorithms to compressed sensing problems to demonstrate the theoretical findings and the enhancements achieved through the proposed framework.

The security proofs of the Quantum Key Distribution (QKD) protocols make certain assumptions about the operations of physical systems. Thus, appropriate modelling of devices to ensure that their operations are consistent with the models assumed in the security proof is imperative. In this paper, we explore the Trojan horse attack (THA) using Measurement Device Independent (MDI) QKD integrated photonic chips and how to avoid some of the security vulnerabilities using only on-chip components. We show that a monitor photodiode paired appropriately with enough optical isolation, given the sensitivity of the photodiode, can detect high power sniffing attacks. We also show that the placement of amplitude modulators with respect to back reflecting components and their switching time can be used to thwart a THA., Comment: 10 pages, 12 figures

Liver cirrhosis, a leading cause of global mortality, requires precise segmentation of ROIs for effective disease monitoring and treatment planning. Existing segmentation models often fail to capture complex feature interactions and generalize across diverse datasets. To address these limitations, we propose a novel synergistic theory that leverages complementary latent spaces for enhanced feature interaction modeling. Our proposed architecture, nnSynergyNet3D integrates continuous and discrete latent spaces for 3D volumes and features auto-configured training. This approach captures both fine-grained and coarse features, enabling effective modeling of intricate feature interactions. We empirically validated nnSynergyNet3D on a private dataset of 628 high-resolution T1 abdominal MRI scans from 339 patients. Our model outperformed the baseline nnUNet3D by approximately 2%. Additionally, zero-shot testing on healthy liver CT scans from the public LiTS dataset demonstrated superior cross-modal generalization capabilities. These results highlight the potential of synergistic latent space models to improve segmentation accuracy and robustness, thereby enhancing clinical workflows by ensuring consistency across CT and MRI modalities.

Recent research focused on single photon emitters (SPEs) hosted by layered semiconductors, particularly hexagonal boron nitride (hBN), has revealed a promising alternative to quantum dots (QDs) for generating single, indistinguishable photons. hBN-based SPEs offer lower material costs, room temperature emission, and easy integration into potential optoelectronic devices due to the layered structure of the host crystal. This work presents compact hybrid electroluminescence devices, in which GaN laser diodes (LDs) are used for bottom-to-top excitation of hBN nanoflakes deposited on the laser facets. Our approach circumvents the issue of electroluminescence generation from hBN and provides access to the SPE's signal without optical driving by an external laser. Using laser diodes upgraded with Bragg reflectors a room-temperature generation of single photons from hBN is confirmed by an 80%-dip in their g(2) second-order correlation. The on-demand emission of single photons at room temperature is demonstrated by driving the laser diodes in pulsed operation, with confidence supported by a measured g(2)(0) value of 0.37.

The calcium isotopes are an ideal system to investigate the evolution of shell structure and magic numbers. Although the properties of surface nucleons in calcium have been well studied, probing the structure of deeply bound nucleons remains a challenge. Here, we report on the first measurement of unbound states in $^{53}$Ca and $^{55}$Ca, populated from \ts{54,56}Ca($p,pn$) reactions at a beam energy of around 216 MeV/nucleon at the RIKEN Radioactive Isotopes Beam Factory. The resonance properties, partial cross sections, and momentum distributions of these unbound states were analyzed. Orbital angular momentum $l$ assignments were extracted from momentum distributions based on calculations using the distorted wave impulse approximation (DWIA) reaction model. The resonances at excitation energies of 5516(41)\,keV in $^{53}$Ca and 6000(250)\,keV in $^{55}$Ca indicate a significant $l$\, =\,3 component, providing the first experimental evidence for the $\nu 0f_{7/2}$ single-particle strength of unbound hole states in the neutron-rich Ca isotopes. The observed excitation energies and cross-sections point towards extremely localized and well separated strength distributions, with some fragmentation for the $\nu 0f_{7/2}$ orbital in $^{55}$Ca. These results are in good agreement with predictions from shell-model calculations using the effective GXPF1Bs interaction and \textit{ab initio} calculations and diverge markedly from the experimental distributions in the nickel isotones at $Z=28$., Comment: 13 pages, 7 figures

A novel method to reconstruct the energy of hadronic showers in the CMS High Granularity Calorimeter (HGCAL) is presented. The HGCAL is a sampling calorimeter with very fine transverse and longitudinal granularity. The active media are silicon sensors and scintillator tiles readout by SiPMs and the absorbers are a combination of lead and Cu/CuW in the electromagnetic section, and steel in the hadronic section. The shower reconstruction method is based on graph neural networks and it makes use of a dynamic reduction network architecture. It is shown that the algorithm is able to capture and mitigate the main effects that normally hinder the reconstruction of hadronic showers using classical reconstruction methods, by compensating for fluctuations in the multiplicity, energy, and spatial distributions of the shower's constituents. The performance of the algorithm is evaluated using test beam data collected in 2018 prototype of the CMS HGCAL accompanied by a section of the CALICE AHCAL prototype. The capability of the method to mitigate the impact of energy leakage from the calorimeter is also demonstrated.

We study the phase transition behavior of the ferroelectric BaTi$_{0.8}$Zr$_{0.2}$O$_3$ in the paraelectric region. The temperature dependencies of thermal, polar, elastic and dielectric properties indicate the presence of local structures above the paraelectric-ferroelectric transition temperature Tc = 292 K. The non-zero remnant polarization is measured up to a characteristic temperature T* ~350 K, which coincides with the temperature where the dielectric constant deviates from Curie-Weiss law. Resonant Piezoelectric Spectroscopy shows that DC field-cooling above Tc using fields smaller than the coercive field leads to an elastic response and remnant piezoelectricity below T*, which likely corresponds to the coherence temperature associated with polar nanostructures in ferroelectrics. The observed remnant effect is attributed to the reorientation of polar nanostructures above Tc., Comment: Accepted version in Applied Physics Letters. 13 pages; 4 figures. (Supp Material in version 1)

Affordances represent the inherent effect and action possibilities that objects offer to the agents within a given context. From a theoretical viewpoint, affordances bridge the gap between effect and action, providing a functional understanding of the connections between the actions of an agent and its environment in terms of the effects it can cause. In this study, we propose a deep neural network model that unifies objects, actions, and effects into a single latent vector in a common latent space that we call the affordance space. Using the affordance space, our system can generate effect trajectories when action and object are given and can generate action trajectories when effect trajectories and objects are given. Our model does not learn the behavior of individual objects acted upon by a single agent. Still, rather, it forms a `shared affordance representation' spanning multiple agents and objects, which we call Affordance Equivalence. Affordance Equivalence facilitates not only action generalization over objects but also Cross Embodiment transfer linking actions of different robots. In addition to the simulation experiments that demonstrate the proposed model's range of capabilities, we also showcase that our model can be used for direct imitation in real-world settings., Comment: 10 pages, 9 figures, Submitted to IEEE Transactions on Cognitive and Developmental Systems

In recent years, Large Language Models (LLMs) have shown remarkable performance in generating human-like text, proving to be a valuable asset across various applications. However, adapting these models to incorporate new, out-of-domain knowledge remains a challenge, particularly for facts and events that occur after the model's knowledge cutoff date. This paper investigates the effectiveness of Supervised Fine-Tuning (SFT) as a method for knowledge injection in LLMs, specifically focusing on the domain of recent sporting events. We compare different dataset generation strategies -- token-based and fact-based scaling -- to create training data that helps the model learn new information. Our experiments on GPT-4 demonstrate that while token-based scaling can lead to improvements in Q&A accuracy, it may not provide uniform coverage of new knowledge. Fact-based scaling, on the other hand, offers a more systematic approach to ensure even coverage across all facts. We present a novel dataset generation process that leads to more effective knowledge ingestion through SFT, and our results show considerable performance improvements in Q&A tasks related to out-of-domain knowledge. This study contributes to the understanding of domain adaptation for LLMs and highlights the potential of SFT in enhancing the factuality of LLM responses in specific knowledge domains., Comment: 16 pages; 7 figures. updated authors list

Distributed systems store data objects redundantly to balance the data access load over multiple nodes. Load balancing performance depends mainly on 1) the level of storage redundancy and 2) the assignment of data objects to storage nodes. We analyze the performance implications of these design choices by considering four practical storage schemes that we refer to as clustering, cyclic, block and random design. We formulate the problem of load balancing as maintaining the load on any node below a given threshold. Regarding the level of redundancy, we find that the desired load balance can be achieved in a system of $n$ nodes only if the replication factor $d = \Omega(\log(n)^{1/3})$, which is a necessary condition for any storage design. For clustering and cyclic designs, $d = \Omega(\log(n))$ is necessary and sufficient. For block and random designs, $d = \Omega(\log(n))$ is sufficient but unnecessary. Whether $d = \Omega(\log(n)^{1/3})$ is sufficient remains open. The assignment of objects to nodes essentially determines which objects share the access capacity on each node. We refer to the number of nodes jointly shared by a set of objects as the \emph{overlap} between those objects. We find that many consistently slight overlaps between the objects (block, random) are better than few but occasionally significant overlaps (clustering, cyclic). However, when the demand is ''skewed beyond a level'' the impact of overlaps becomes the opposite. We derive our results by connecting the load-balancing problem to mathematical constructs that have been used to study other problems. For a class of storage designs containing the clustering and cyclic design, we express load balance in terms of the maximum of moving sums of i.i.d. random variables, which is known as the scan statistic. For random design, we express load balance by using the occupancy metric for random allocation with complexes., Comment: Submitted to Transactions on Networking