Hasil untuk "Semantics"

Yujun Shen, Jinjin Gu, Xiaoou Tang et al.

Despite the recent advance of Generative Adversarial Networks (GANs) in high-fidelity image synthesis, there lacks enough understanding of how GANs are able to map a latent code sampled from a random distribution to a photo-realistic image. Previous work assumes the latent space learned by GANs follows a distributed representation but observes the vector arithmetic phenomenon. In this work, we propose a novel framework, called InterFaceGAN, for semantic face editing by interpreting the latent semantics learned by GANs. In this framework, we conduct a detailed study on how different semantics are encoded in the latent space of GANs for face synthesis. We find that the latent code of well-trained generative models actually learns a disentangled representation after linear transformations. We explore the disentanglement between various semantics and manage to decouple some entangled semantics with subspace projection, leading to more precise control of facial attributes. Besides manipulating gender, age, expression, and the presence of eyeglasses, we can even vary the face pose as well as fix the artifacts accidentally generated by GAN models. The proposed method is further applied to achieve real image manipulation when combined with GAN inversion methods or some encoder-involved models. Extensive results suggest that learning to synthesize faces spontaneously brings a disentangled and controllable facial attribute representation.

David Bau, Bolei Zhou, A. Khosla et al.

We propose a general framework called Network Dissection for quantifying the interpretability of latent representations of CNNs by evaluating the alignment between individual hidden units and a set of semantic concepts. Given any CNN model, the proposed method draws on a data set of concepts to score the semantics of hidden units at each intermediate convolutional layer. The units with semantics are labeled across a broad range of visual concepts including objects, parts, scenes, textures, materials, and colors. We use the proposed method to test the hypothesis that interpretability is an axis-independent property of the representation space, then we apply the method to compare the latent representations of various networks when trained to solve different classification problems. We further analyze the effect of training iterations, compare networks trained with different initializations, and measure the effect of dropout and batch normalization on the interpretability of deep visual representations. We demonstrate that the proposed method can shed light on characteristics of CNN models and training methods that go beyond measurements of their discriminative power.

Marco Baroni, Georgiana Dinu, Germán Kruszewski

Context-predicting models (more commonly known as embeddings or neural language models) are the new kids on the distributional semantics block. Despite the buzz surrounding these models, the literature is still lacking a systematic comparison of the predictive models with classic, count-vector-based distributional semantic approaches. In this paper, we perform such an extensive evaluation, on a wide range of lexical semantics tasks and across many parameter settings. The results, to our own surprise, show that the buzz is fully justified, as the context-predicting models obtain a thorough and resounding victory against their count-based counterparts.

M. Herlihy, Jeannette M. Wing

E. Emerson

F. Bancilhon, N. Spyratos

Xiaoyu Zhou, Yaoshuai Dang, Jinling Song et al.

Crop yield estimation, particularly early-season yield prediction, is highly important for global food security and disaster mitigation. In this study, we utilized deep learning models combined with remote sensing data to develop in-season crop yield estimation models, enabling immediate yield prediction. We employed a convolutional neural network (CNN) for spatial feature extraction and a long short-term memory network (LSTM) for temporal patterns, complemented by Gaussian process regression (GP) that introduced geographical coordinates. Three groups of in-season yield prediction experiments were designed, utilizing four-phase, two-phase, and single-phase data, respectively. The results indicated that under the two-phase training scheme, the LSTM_GP model achieved the highest performance in the sixth period, with an <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msup><mrow><mi>R</mi></mrow><mrow><mn>2</mn></mrow></msup></mrow></semantics></math></inline-formula> value of 0.61 and a root mean square error (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mi>R</mi><mi>M</mi><mi>S</mi><mi>E</mi></mrow></semantics></math></inline-formula>) value of 983.38 kg/ha. When trained on single-phase data at the twelfth phase (approximately mid-to-late July), the LSTM_GP model also performed best, attaining an <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msup><mrow><mi>R</mi></mrow><mrow><mn>2</mn></mrow></msup></mrow></semantics></math></inline-formula> value of 0.62 and an <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mi>R</mi><mi>M</mi><mi>S</mi><mi>E</mi></mrow></semantics></math></inline-formula> value of 969.06 kg/ha. The single-phase prediction model outperformed time-series models in yield prediction accuracy. The periods from mid-to-late July to early-to-mid August represent critical crop growth stages were essential for accurate yield prediction. From our research, we found that adding GP can improve the prediction accuracy, especially for LSTM. Moreover, the proposed single-phase prediction model realized reliable crop yield prediction as well as the silking to early grain-filling stage (mid-to-late July), providing a critical lead time of approximately 2–2.5 months before harvest to support pre-harvest agricultural decision-making.

Ewan Klein

Jingqian Wu, Yunbo Jia, Shengpeng Xu et al.

Event cameras provide a promising sensing modality for high-speed and high-dynamic-range vision by asynchronously capturing brightness changes. A fundamental task in event-based vision is event-to-video (E2V) reconstruction, which aims to recover intensity videos from event streams. Most existing E2V approaches formulate reconstruction as a temporal--spatial signal recovery problem, relying on temporal aggregation and spatial feature learning to infer intensity frames. While effective to some extent, this formulation overlooks a critical limitation of event data: due to the change-driven sensing mechanism, event streams are inherently semantically under-determined, lacking object-level structure and contextual information that are essential for faithful reconstruction. In this work, we revisit E2V from a semantic perspective and argue that effective reconstruction requires going beyond temporal and spatial modeling to explicitly account for missing semantic information. Based on this insight, we propose \textit{Semantic-E2VID}, a semantic-enriched end-to-end E2V framework that reformulates reconstruction as a process of semantic learning, fusing and decoding. Our approach first performs semantic abstraction by bridging event representations with semantics extracted from a pretrained Segment Anything Model (SAM), while avoiding modality-induced feature drift. The learned semantics are then fused into the event latent space in a representation-compatible manner, enabling event features to capture object-level structure and contextual cues. Furthermore, semantic-aware supervision is introduced to explicitly guide the reconstruction process toward semantically meaningful regions, complementing conventional pixel-level and temporal objectives. Extensive experiments on six public benchmarks demonstrate that Semantic-E2VID consistently outperforms state-of-the-art E2V methods.

Yll Buzoku, David. J. Pym

The proof theory and semantics of intuitionistic modal logics have been studied by Simpson in terms of Prawitz-style labelled natural deduction systems and Kripke models. An alternative to model-theoretic semantics is provided by proof-theoretic semantics, which is a logical realization of inferentialism, in which the meaning of constructs is understood through their use. The key idea in proof-theoretic semantics is that of a base of atomic rules, all of which refer only to propositional atoms and involve no logical connectives. A specific form of proof-theoretic semantics, known as base-extension semantics (B-eS), is concerned with the validity of formulae and provides a direct counterpart to Kripke models that is grounded in the provability of atomic formulae in a base. We establish, systematically, B-eS for Simpson's intuitionistic modal logics and, also systematically, obtain soundness and completeness theorems with respect to Simpson's natural deduction systems.

José Luis Díaz Palencia

Cosmic rays exhibit anomalous diffusion behaviors in the heliospheric environment that cannot be adequately described by classical diffusion models. In this paper, we develop a theoretical framework employing a fractional Fokker–Planck equation to model the anomalous transport of cosmic rays. This approach accounts for the observed non-Gaussian distributions, long-range correlations and memory effects in cosmic ray fluxes. We derive analytical solutions using the Adomian Decomposition Method and express them in terms of Mittag-Leffler functions and Lévy stable distributions. The model parameters, including the fractional orders <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>α</mi></semantics></math></inline-formula> and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>μ</mi></semantics></math></inline-formula> and the entropic index <i>q</i>, are estimated by a short comparison between theoretical predictions and observational data from cosmic ray experiments. Our findings suggest that the integration of fractional calculus and non-extensive statistics can be employed for describing the cosmic ray propagation and the anomalous diffusion observed in the heliosphere.

Yu-An Huang, Xiyue Cao, Zhu-Hong You et al.

Abstract The rise of single-cell sequencing has revolutionized the exploration of drug resistance, revealing the crucial role of cellular heterogeneity in advancing precision medicine. By building computational models from existing single-cell drug response data, we can rapidly annotate cellular responses to drugs. To this end, we developed scGSDR, integrating two computational pipelines grounded in the knowledge of cellular states and gene signaling pathways, both essential for understanding biological gene semantics. scGSDR enhances predictive performance by incorporating gene semantics and employs an interpretability module to identify pathways contributing to drug resistance phenotypes. Extensive validation demonstrates scGSDR’s superior predictive accuracy when trained with bulk RNA-seq or scRNA-seq data. The model’s application has extended from single-drug predictions to scenarios involving drug combinations. Leveraging pathways of known drug target genes, we found that scGSDR’s cell-pathway attention scores are biologically interpretable, which helped us identify potential drug-related genes. Literature review of top-ranking genes in predictions such as BCL2, CCND1, and PIK3CA for PLX4720 confirmed their relevance. Overall, scGSDR, by incorporating gene semantics, enhances predictive modeling of cellular responses to diverse drugs, proving invaluable for scenarios involving both single drug and combination therapies and effectively identifying key resistance-related pathways, thus advancing precision medicine and targeted therapy development.

Hang Zhang, Zimo Gong, Chen Hu et al.

This study focuses on pest and disease detection in cistanche, proposing a Transformer-based object detection network enhanced by a bridging attention mechanism and bridging loss function, demonstrating outstanding performance in complex agricultural scenarios. The bridging attention mechanism dynamically fuses low-level details and high-level semantics, significantly improving detection capabilities for small targets and complex backgrounds. Experimental results show that the method achieves an average accuracy of 0.93, a precision of 0.95, a recall of 0.92, and mAP@50 and mAP@75 scores of 0.92 and 0.90, outperforming traditional self-attention mechanisms and CBAM modules. These results confirm the method’s ability to overcome challenges such as unclear disease features and small target sizes, providing robust support for precision pest detection. The research contributes to smart agricultural disease management and the sustainable development of cistanche cultivation while laying a solid foundation for future agricultural intelligence applications.

M. Fitting

The use of conventional classical logic is misleading for characterizing the behavior of logic programs because a logic program, when queried, will do one of three things: succeed with the query, fail with it, or not respond because it has fallen into infinite backtracking. In [7] Kleene proposed a three-valued logic for use in recursive function theory. The so-called third truth value was really undefined: truth value not determined. This logic is a useful tool in logic-program specification, and in particular, for describing models. (See [11].) Tarski showed that formal languages, like arithmetic, cannot contain their own truth predicate because one could then construct a paradoxical sentence that effectively asserts its own falsehood. Natural languages do allow the use of "is true", so by Tarski's argument a semantics for natural language must leave truth-value gaps: some sentences must fail to have a truth value. In [8] Kripke showed how a model having truth-value gaps, using Kleene's three-valued logic, could be specified. The mechanism he used is a famiUar one in program semantics: consider the least fixed point of a certain monotone operator. But that operator must be defined on a space involving three-valued logic, and for Kripke's application it will not be continuous. We apply techniques similar to Kripke's to logic programs. We associate with each program a monotone operator on a space of three-valued logic interpretations, or better partial interpretations. This space is not a complete lattice, and the operators are not, in general, continuous. But least and other fixed points do exist. These fixed points are shown to provide suitable three-valued program models. They relate closely to the least and greatest fixed points of the operators used in [1]. Because of the extra machinery involved, our treatment allows for a natural consideration of negation, and indeed, of the other prepositional connectives as well. And because of the elaborate structure of fixed points available, we are able to

Assaf Libman, Nir Oren, Bruno Yun

Weighted gradual semantics provide an acceptability degree to each argument representing the strength of the argument, computed based on factors including background evidence for the argument, and taking into account interactions between this argument and others. We introduce four important problems linking gradual semantics and acceptability degrees. First, we reexamine the inverse problem, seeking to identify the argument weights of the argumentation framework which lead to a specific final acceptability degree. Second, we ask whether the function mapping between argument weights and acceptability degrees is injective or a homeomorphism onto its image. Third, we ask whether argument weights can be found when preferences, rather than acceptability degrees for arguments are considered. Fourth, we consider the topology of the space of valid acceptability degrees, asking whether "gaps" exist in this space. While different gradual semantics have been proposed in the literature, in this paper, we identify a large family of weighted gradual semantics, called abstract weighted based gradual semantics. These generalise many of the existing semantics while maintaining desirable properties such as convergence to a unique fixed point. We also show that a sub-family of the weighted gradual semantics, called abstract weighted (L^p,λ,μ)-based gradual semantics and which include well-known semantics, solve all four of the aforementioned problems.

Lydia Blümel, Matthias Thimm

We consider the notion of a vacuous reduct semantics for abstract argumentation frameworks, which, given two abstract argumentation semantics σ and τ, refines σ (base condition) by accepting only those σ-extensions that have no non-empty τ-extension in their reduct (vacuity condition). We give a systematic overview on vacuous reduct semantics resulting from combining different admissibility-based and conflict-free semantics and present a principle-based analysis of vacuous reduct semantics in general. We provide criteria for the inheritance of principle satisfaction by a vacuous reduct semantics from its base and vacuity condition for established as well as recently introduced principles in the context of weak argumentation semantics. We also conduct a principle-based analysis for the special case of undisputed semantics.

Qicheng Zeng, Zhaojun Nan, Sheng Zhou

Coded computing is recognized as a promising solution to address the privacy leakage problem and the straggling effect in distributed computing. This technique leverages coding theory to recover computation tasks using results from a subset of workers. In this paper, we propose the adaptive privacy-preserving coded computing (APCC) strategy, designed to be applicable to various types of computation tasks, including polynomial and non-polynomial functions, and to adaptively provide accurate or approximated results. We prove the optimality of APCC in terms of encoding rate, defined as the ratio between the computation loads of tasks before and after encoding, based on the optimal recovery threshold of Lagrange Coded Computing. We demonstrate that APCC guarantees information-theoretical data privacy preservation. Mitigation of the straggling effect in APCC is achieved through hierarchical task partitioning and task cancellation, which further reduces computation delays by enabling straggling workers to return partial results of assigned tasks, compared to conventional coded computing strategies. The hierarchical task partitioning problems are formulated as mixed-integer nonlinear programming (MINLP) problems with the objective of minimizing task completion delay. We propose a low-complexity maximum value descent (MVD) algorithm to optimally solve these problems. The simulation results show that APCC can reduce the task completion delay by a range of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>20.3</mn><mo>%</mo></mrow></semantics></math></inline-formula> to <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>47.5</mn><mo>%</mo></mrow></semantics></math></inline-formula> when compared to other state-of-the-art benchmarks.

Haodong Zhang, ZhiKe Chen, Haocheng Xu et al.

Capturing and preserving motion semantics is essential to motion retargeting between animation characters. However, most of the previous works neglect the semantic information or rely on human-designed joint-level representations. Here, we present a novel Semantics-aware Motion reTargeting (SMT) method with the advantage of vision-language models to extract and maintain meaningful motion semantics. We utilize a differentiable module to render 3D motions. Then the high-level motion semantics are incorporated into the motion retargeting process by feeding the vision-language model with the rendered images and aligning the extracted semantic embeddings. To ensure the preservation of fine-grained motion details and high-level semantics, we adopt a two-stage pipeline consisting of skeleton-aware pre-training and fine-tuning with semantics and geometry constraints. Experimental results show the effectiveness of the proposed method in producing high-quality motion retargeting results while accurately preserving motion semantics.

Beishui Liao, Leendert van der Torre

In this paper we introduce a novel semantics, called defense semantics, for Dung's abstract argumentation frameworks in terms of a notion of (partial) defence, which is a triple encoding that one argument is (partially) defended by another argument via attacking the attacker of the first argument. In terms of defense semantics, we show that defenses related to self-attacked arguments and arguments in 3-cycles are unsatifiable under any situation and therefore can be removed without affecting the defense semantics of an AF. Then, we introduce a new notion of defense equivalence of AFs, and compare defense equivalence with standard equivalence and strong equivalence, respectively. Finally, by exploiting defense semantics, we define two kinds of reasons for accepting arguments, i.e., direct reasons and root reasons, and a notion of root equivalence of AFs that can be used in argumentation summarization.

Erik Voogd, Einar Broch Johnsen, Alexandra Silva et al.

We present a new symbolic execution semantics of probabilistic programs that include observe statements and sampling from continuous distributions. Building on Kozen's seminal work, this symbolic semantics consists of a countable collection of measurable functions, along with a partition of the state space. We use the new semantics to provide a full correctness proof of symbolic execution for probabilistic programs. We also implement this semantics in the tool symProb, and illustrate its use on examples.