Abstract : We introduce the DET Curve as a means of representing performance on detection tasks that involve a tradeoff of error types. We discuss why we prefer it to the traditional ROC Curve and offer several examples of its use in speaker recognition and language recognition. We explain why it is likely to produce approximately linear curves. We also note special points that may be included on these curves, how they are used with multiple targets, and possible further applications.
Global context information is particularly important for comprehensive scene understanding. It helps clarify local confusions and smooth predictions to achieve fine-grained and coherent results. However, most existing light field processing methods leverage convolution layers to model spatial and angular information. The limited receptive field restricts them to learn long-range dependency in LF structure. In this article, we propose a novel network based on deep efficient transformers (i.e., LF-DET) for LF spatial super-resolution. It develops a spatial-angular separable transformer encoder with two modeling strategies termed as sub-sampling spatial modeling and multi-scale angular modeling for global context interaction. Specifically, the former utilizes a sub-sampling convolution layer to alleviate the problem of huge computational cost when capturing spatial information within each sub-aperture image. In this way, our model can cascade more transformers to continuously enhance feature representation with limited resources. The latter processes multi-scale macro-pixel regions to extract and aggregate angular features focusing on different disparity ranges to well adapt to disparity variations. Besides, we capture strong similarities among surrounding pixels by dynamic positional encodings to fill the gap of transformers that lack of local information interaction. The experimental results on both real-world and synthetic LF datasets confirm our LF-DET achieves a significant performance improvement compared with state-of-the-art methods. Furthermore, our LF-DET shows high robustness to disparity variations through the proposed multi-scale angular modeling.
Deepfake detection is a long-established research topic vital for mitigating the spread of malicious misinformation. Unlike prior methods that provide either binary classification results or textual explanations separately, we introduce a novel method capable of generating both simultaneously. Our method harnesses the multi-modal learning capability of the pre-trained CLIP and the unprecedented interpretability of large language models (LLMs) to enhance both the generalization and explainability of deep-fake detection. Specifically, we introduce a multi-modal face forgery detector (M2F2-Det) that employs tailored face forgery prompt learning, incorporating the pre-trained CLIP to improve generalization to unseen forgeries. Also, M2F2-Det incorporates an LLM to provide detailed textual explanations of its detection decisions, enhancing interpretability by bridging the gap between natural language and subtle cues of facial forgeries. Empirically, we evaluate M2F2-Det on both detection and explanation generation tasks, where it achieves state-of-the-art performance, demonstrating its effectiveness in identifying and explaining diverse forgeries. Source code is available at $\color{magenta}{link}$.
To improve the precision of defect categorization and localization in images, this paper proposes an approach for detecting surface defects in hot-rolled steel strips. The approach uses an improved YOLOv5 network model to overcome the issues of inadequate feature extraction capacity and suboptimal feature integration when identifying surface defects on steel strips. The proposed method achieves higher detection accuracy and localization precision, making it more competitive and applicable in real production. Firstly, the multi-scale feature fusion (MSF) strategy is utilized to fuse shallow and deep features effectively and enrich detailed information relevant to target defects. Secondly, the CSPLayer Res2Attention block (CRA block) residual module is introduced to reduce the loss of defect information during hierarchical transmission, thereby enhancing the extraction of fine-grained features and improving the perception of details and global features. Finally, the experimental results indicate that the mAP on the NEU-DET and GC10-DET datasets approaches 78.5% and 67.3%, respectively, which is 4.9% and 2.1% higher than that of the baseline. Meanwhile, it has higher precision and more precise localization capabilities than other methods. Furthermore, it also achieves 59.2% mAP on the APDDD dataset, indicating its potential for growth in further domains.
Large numbers of synthesized videos from diffusion models pose threats to information security and authenticity, leading to an increasing demand for generated content detection. However, existing video-level detection algorithms primarily focus on detecting facial forgeries and often fail to identify diffusion-generated content with a diverse range of semantics. To advance the field of video forensics, we propose an innovative algorithm named Multi-Modal Detection(MM-Det) for detecting diffusion-generated videos. MM-Det utilizes the profound perceptual and comprehensive abilities of Large Multi-modal Models (LMMs) by generating a Multi-Modal Forgery Representation (MMFR) from LMM's multi-modal space, enhancing its ability to detect unseen forgery content. Besides, MM-Det leverages an In-and-Across Frame Attention (IAFA) mechanism for feature augmentation in the spatio-temporal domain. A dynamic fusion strategy helps refine forgery representations for the fusion. Moreover, we construct a comprehensive diffusion video dataset, called Diffusion Video Forensics (DVF), across a wide range of forgery videos. MM-Det achieves state-of-the-art performance in DVF, demonstrating the effectiveness of our algorithm. Both source code and DVF are available at https://github.com/SparkleXFantasy/MM-Det.
Digital extension tools (DETs) include phone calls, WhatsApp groups and specialised smartphone applications used for agricultural knowledge brokering. We researched processes through which DETs have (and have not) been used by farmers and other extension actors in low- and middle-income countries. We interviewed 40 DET developers across 21 countries and 101 DET users in Bihar, India. We found DET use is commonly constrained by fifteen pitfalls (unawareness of DET, inaccessible device, inaccessible electricity, inaccessible mobile network, insensitive to digital illiteracy, insensitive to illiteracy, unfamiliar language, slow to access, hard to interpret, unengaging, insensitive to user's knowledge, insensitive to priorities, insensitive to socio-economic constraints, irrelevant to farm, distrust). These pitfalls partially explain why women, less educated and less wealthy farmers often use DETs less, as well as why user-driven DETs (e.g. phone calls and chat apps) are often used more than externally-driven DETs (e.g. specialised smartphone apps). Our second key finding was that users often made - not just found - DETs useful for themselves and others. This suggests the word ‘appropriation’ conceptualises DET use more accurately and helpfully than the word ‘adoption’. Our final key finding was that developers and users advocated almost ubiquitously for involving desired users in DET provision. We synthesise these findings in a one-page framework to help funders and developers facilitate more useable, useful and positively impactful DETs. Overall, we conclude developers increase DET use by recognizing users as fellow developers – either through collaborative design or by designing adaptable DETs that create room for user innovation.
This paper presents a formulation alongside a numerical solution algorithm to describe the mechanical response of bodies made of a large class of viscoelastic materials undergoing arbitrary quasistatic finite deformations. With the objective of having a unified formulation that applies to a wide range of highly compressible, nearly incompressible, and fully incompressible soft organic materials in a numerically tractable manner, the viscoelasticity is described within a Lagrangian setting by a two-potential mixed formulation. In this formulation, the deformation field, a pressure field that ensues from a Legendre transform, and an internal variable of state Fv that describes the viscous part of the deformation are the independent fields. Consistent with the experimental evidence that viscous deformation is a volume-preserving process, the internal variable Fv is required to satisfy the constraint det Fv=1. To solve the resulting initial-boundary-value problem, a numerical solution algorithm is proposed that is based on a finite-element (FE) discretization of space and a finite-difference discretization of time. Specifically, a Variational Multiscale FE method is employed that allows for an arbitrary combination of shape functions for the deformation and pressure fields. To deal with the challenging non-convex constraint det Fv=1, a new time integration scheme is introduced that allows to convert any explicit or implicit scheme of choice into a stable scheme that preserves the constraint det Fv=1 identically. A series of test cases is presented that showcase the capabilities of the proposed formulation.
Cecilie Marie Nielsen, Pernille Kølbæk, David Dines
et al.
AbstractRatings on the Positive and Negative Syndrome Scale (PANSS) are ideally based on both a patient interview and an informant questionnaire. In research and clinical settings, however, the informant questionnaire is often omitted. This study investigated the consequences of omitting informant information by comparing PANSS ratings of patients with schizophrenia (n = 49 patients, 77 ratings) conducted with and without informant information, respectively. Additionally, changes in symptom severity over time based on ratings with and without informant information were also compared for the full PANSS and the six-item version of the PANSS (PANSS-6). PANSS ratings including informant information were higher than those without, both at the total score and individual item level. Additionally, the full PANSS appeared less “responsive” to baseline-to-endpoint changes for ratings without informant information compared to ratings including informant information, while no differences were found for the PANSS-6.
Detta kapitel belyser fundamentalantagandena om samhället och tillvaron i Vetenskapsakademiens politiska ideologi. Det första avsnittet analyserar presidietalens redogörelser för förhållandet mellan den mänskliga samlevnaden och naturen samt deras föreställningar om samhällets ursprung, bevekelsegrunder och utveckling. Det andra avsnittet behandlar framställningarna av samhällsgemenskapens organisering och bindemedel, det tredje de organiska, mekaniska och arkitektoniska bilder som användes för att åskådliggöra den.
Determining inorganic nutrient profiles to support understanding of nitrogen transformations in stream sediments is challenging, due to nitrification and denitrification being confined to particular conditions in potentially heterogeneous sediment influenced by benthic microalgae, rooted aquatic plants and/or diel light cycles. The diffusive gradients in thin films (DGT) and diffusive equilibration in thin films (DET) techniques allow in situ determination of porewater concentration profiles, and distributions for some solutes. In this study, DGT, DET and conventional porewater extraction (sectioning and centrifugation) methods were compared for ammonium and nitrate in stream sediments under light and dark conditions. Two-dimensional distributions of Fe(ii) and PO4-P were also provided to indicate the degree of spatial and temporal heterogeneity in sediment porewater, which can explain the sources and sinks of ammonium at various depths in the sediments. Although the conventional porewater extraction method consistently measured higher NH4-N concentrations than the DGT and DET techniques, the study showed that the DET measurements were the most reliable indicator of porewater NH4-N concentrations, with the DGT data being usefully supplementary. However, a large proportion of the NO3-N concentrations measured by DGT and DET were close to or below the method detection limits. Therefore, further development of these techniques is required to reduce the blanks and detection limits to allow natural low sediment porewater NO3-N concentrations to be accurately monitored using DGT and DET. The study indicated that benthic microalgae had direct and indirect influences on porewater nutrient distributions over light-dark cycles. Overall, DGT and DET techniques can be useful for monitoring porewater nutrient concentrations and profiles and for determining how biological processes drive changes in sediment nutrient concentrations and distributions.
Laura Søvsø Thomasen: Tracing Ørsted – Interdisciplinary Research and Dissemination Perspectives from the Royal Danish Library’s Hans Christian Ørsted Archive 2020 marks the 200th anniversary for Hans Christian Ørsted’s discovery of electromagnetism. In connection with the celebration of the bicentennial, the Royal Danish Library will publish online their entire Ørsted archive, which includes a substantial collection of letters to and from Ørsted, a large number of scientific papers on physics, chemistry and mathematics, as well as a plethora of different works on everything from language over politics to aesthetics. With the digitalisation of the archive, the library has created a number of teaching materials available for students in the Danish upper secondary school that showcase the interdisciplinary methods and works by Ørsted but also emphasises the interdisciplinary work required by the students to solve the problem sets. In the article, I explore how the digitalisation of the Ørsted archive opens up for new perspectives both in academic research as well as didactic perspectives in relation to the teaching materials accompanying the archive. In terms of new research perspective, the Ørsted archive showcases how Ørsted not only was interested in a variety of subjects in and around both the natural and cultural sciences, but that he also was truly interdisciplinary himself. Through source material from the archive I show how when approached from the interdisciplinary fields ‘Literature and Science’ and ‘Visual Culture of Science’ that Ørsted integrated both illustrations and literary components to communicate his scientific arguments.
The log-det distance between two aligned DNA sequences was introduced as a tool for statistically consistent inference of a gene tree under simple non-mixture models of sequence evolution. Here we prove that the log-det distance, coupled with a distance-based tree construction method, also permits consistent inference of species trees under mixture models appropriate to aligned genomic-scale sequences data. Data may include sites from many genetic loci, which evolved on different gene trees due to incomplete lineage sorting on an ultrametric species tree, with different time-reversible substitution processes. The simplicity and speed of distance-based inference suggests log-det based methods should serve as benchmarks for judging more elaborate and computationally-intensive species trees inference methods.
The Vol‐Det Conjecture relates the volume and the determinant of a hyperbolic alternating link in S3 . We use exact computations of Mahler measures of two‐variable polynomials to prove the Vol‐Det Conjecture for many infinite families of alternating links.
The Alpha-Beta Log-Det divergences for positive definite matrices are flexible divergences that are parameterized by two real constants and are able to specialize several relevant classical cases like the squared Riemannian metric, the Steins loss, the S-divergence, etc. A novel classification criterion based on these divergences is optimized to address the problem of classification of the motor imagery movements. This research paper is divided into three main sections in order to address the above mentioned problem: (1) Firstly, it is proven that a suitable scaling of the class conditional covariance matrices can be used to link the Common Spatial Pattern (CSP) solution with a predefined number of spatial filters for each class and its representation as a divergence optimization problem by making their different filter selection policies compatible; (2) A closed form formula for the gradient of the Alpha-Beta Log-Det divergences is derived that allows to perform optimization as well as easily use it in many practical applications; (3) Finally, in similarity with the work of Samek et al. 2014, which proposed the robust spatial filtering of the motor imagery movements based on the beta-divergence, the optimization of the Alpha-Beta Log-Det divergences is applied to this problem. The resulting subspace algorithm provides a unified framework for testing the performance and robustness of the several divergences in different scenarios.
Wireless control system for industrial automation has been gaining increasing popularity in recent years thanks to their ease of deployment and the low cost of their components. However, traditional low sample rate industrial wireless sensor networks cannot support high-speed application, while high-speed IEEE 802.11 networks are not designed for real-time application and not able to provide deterministic feature. Thus, in this paper, we propose Det-WiFi, a real-time TDMA MAC implementation for high-speed multihop industrial application. It is able to support high-speed applications and provide deterministic network features since it combines the advantages of high-speed IEEE802.11 physical layer and a software Time Division Multiple Access (TDMA) based MAC layer. We implement Det-WiFi on commercial off-the-shelf hardware and compare the deterministic performance between 802.11s and Det-WiFi under the real industrial environment, which is full of field devices and industrial equipment. We changed the hop number and the packet payload size in each experiment, and all of the results show that Det-WiFi has better deterministic performance.