Semantics (Computer Science)

The paper reports on activities carried within the Agrisemantics Working Group of the Research Data Alliance (RDA). The group investigated on what are the current problems research and practitioners experience in their work with semantic resources for agricultural data and elaborated the list of requirements that are the object of this paper. The main findings include the need to broaden the usability of tools so as to make them useful and available to the variety of profiles usually involved in working with semantics resources; the need to online platform to lift users from the burden of local installation; and the need for services that can be integrated in workflows. We further analyze requirements concerning the tools and services and provide details about the process followed to gather evidence from the community.

The paper reports on activities carried within the Agrisemantics Working Group of the Research Data Alliance (RDA). The group investigated on what are the current problems research and practitioners experience in their work with semantic resources for agricultural data, and elaborated the list of requirements that are the object of this paper. The main findings include the need to broaden the usability of tools so as to make them useful and available to the variety of profiles usually involved in working with semantics resources; the need to online platform to lift users from the burden of local installation; and the need for services that can be integrated in workflows. We further analyze requirements concerning the tools and services and provide details about the process followed to gather evidence from the community.

The mathematics of probabilistic generative modelling: VAEs, normalising flows, GANs, and score-based diffusion processes. Concludes with the formal foundations of RLHF, reward modelling, and alignment techniques including PPO applied to language model fine-tuning.

The use of Domain-Specific Languages (DSLs) is a promising field for the development of tools tailored to specific problem spaces, effectively diminishing the complexity of hand-made software. With the goal of making models as precise, simple and reusable as possible, we augment DSLs with concepts from multilevel modelling, where the number of abstraction levels are not limited. This is particularly useful for DSL definitions with behaviour, whose concepts inherently belong to different levels of abstraction. Here, models can represent the state of the modelled system and evolve using model transformations. These transformations can benefit from a multilevel setting, becoming a precise and reusable definition of the semantics for behavioural modelling languages. We present in this paper the concept of Multilevel Coupled Model Transformations, together with examples, formal definitions and tools to assess their conceptual soundness and practical value.

Systems based on statistical and machine learning methods have been shown to be extremely effective and scalable for the analysis of large amount of textual data. However, in the recent years, it becomes evident that one of the most important directions of improvement in natural language processing (NLP) tasks, like word sense disambiguation, coreference resolution, relation extraction, and other tasks related to knowledge extraction, is by exploiting semantics. While in the past, the unavailability of rich and complete semantic descriptions constituted a serious limitation of their applicability, nowadays, the Semantic Web made available a large amount of logically encoded information (e.g. ontologies, RDF(S)-data, linked data, etc.), which constitutes a valuable source of semantics. However, web semantics cannot be easily plugged into machine learning systems. Therefore the objective of this paper is to define a reference methodology for combining semantic information available in the web under the form of logical theories, with statistical methods for NLP. The major problems that we have to solve to implement our methodology concern (i) the selection of the correct and minimal knowledge among the large amount available in the web, (ii) the representation of uncertain knowledge, and (iii) the resolution and the encoding of the rules that combine knowledge retrieved from Semantic Web sources with semantics in the text. In order to evaluate the appropriateness of our approach, we present an application of the methodology to the problem of intra-document coreference resolution, and we show by means of some experiments on the standard dataset, how the injection of knowledge leads to the improvement of this task performance.

Compared to other existing semantic role repositories, FrameNet is characterized by an extremely high number of roles or Frame Elements (FEs), which amount to 8,884 in the last resource release. This represents an interesting issue to investigate both from a theoretical and a practical point of view. In this paper, we analyze the semantics of frame elements by automatically assigning them a set of synsets characterizing the typical FE fillers. We show that the synset repository created for each FE can adequately generalize over the fillers, while providing more informative sense labels than just one generic semantic type. We also evaluate the impact of the enriched FE information on a semantic role labeling task, showing that it can improve classification precision, though at the cost of lower recall.

This paper begins with the premise that adverbs are neglected in computational linguistics. This view derives from two analyses: a literature review and a novel adverb dataset to probe a stateof-the-art language model, thereby uncovering systematic gaps in accounts for adverb meaning. We suggest that using Frame Semantics for characterizing word meaning, as in FrameNet, provides a promising approach to adverb analysis, given its ability to describe ambiguity, semantic roles, and null instantiation.

Program transformation has gained a wide interest since it is used for several purposes: altering semantics of a program, adding features to a program or performing optimizations. In this paper we focus on program transformations at the bytecode level. Because these transformations may introduce errors, our goal is to provide a formal way to verify the update and establish its correctness. The formal framework presented includes a definition of a formal semantics of updates which is the base of a static verification and a scheme based on Hoare triples and weakest precondition calculus to reason about behavioral aspects in bytecode transformation. **Notations: • Effect STK ( a,b): represents the effect of the instruction a on the stack a. • Effect F(a,b): represents the effect on the instruction a on F.

A major challenge in distributed groupware is enabling efficient collaboration by assuring effective communication between the users. Different tools and widgets are commonly used to provide a sense of awareness of other users actions. The problem is compounded when collaboration takes place in a heterogeneous environment, with representation and semantics adapted to computing device's capabilities.

One of the fundamental problems of the RRG Linking Algorithm from syntax to semantics is that the selection of the appropriate syntactic template for an input sentence is often subjected to several ambiguities. In this paper, we examine PP attachment ambiguities focusing on the French verb parler. Within the first section, we deliver an analysis of the possible syntactic templates occurring with parler. Based upon a quantitative evaluation of parler sentences, we then propose a statisticallydriven algorithm for the distinction between core arguments and adjuncts which results in the selection of the most likely template.

According to a connectionist view, mental states consist of the activations of neural units in a connectionist network. We consider the similarity of representations that emerge in unsupervised, self-organization process of neural lattices when exposed to color spectrum stimuli. Self-Organizing Maps (SOM) are trained with color spectrum input, using various vectorial encodings for representation of the input. Further, the SOM is used for heteroassociative mapping to associate color spectrum with color names. Recall of association between the spectra and colors is assessed. It shows that the SOM learns representations for both stimuli and color names, and is able to associate them successfully. The resulting organization is compared through correlation of the activation patterns of the neural maps when responding to color spectrum stimuli. Experiments show that the emerged representations for stimuli are similar with respect to the partitioning-of-activation-space measure almost independently of the encoding used for input representation. This adds a new example in favour of the usability of the state space semantics.

According to a connectionist view, mental states consist of the activations of neural units in a connectionist network. We consider the similarity of representations that emerge in unsupervised, self-organization process of neural lattices when exposed to color spectrum stimuli. Self-Organizing Maps (SOM) are trained with color spectrum input, using various vectorial encodings for representation of the input. Further, the SOM is used for heteroassociative mapping to associate color spectrum with color names. Recall of association between the spectra and colors is assessed. It shows that the SOM learns representations for both stimuli and color names, and is able to associate them successfully. The resulting organization is compared through correlation of the activation patterns of the neural maps when responding to color spectrum stimuli. Experiments show that the emerged representations for stimuli are similar with respect to the partitioning-of-activation-space measure almost independently of the encoding used for input representation. This adds a new example in favour of the usability of the state space semantics.

How we remember the things that we see can be shaped by our prior experiences. Here, we examine how linguistic and sensory experiences interact to influence visual memory. Objects in a visual search that shared phonology (cat-cast) or semantics (dog-fox) with a target were later remembered better than unrelated items. Phonological overlap had a greater influence on memory when targets were cued by spoken words, while semantic overlap had a greater effect when targets were cued by characteristic sounds. The influence of overlap on memory varied as a function of individual differences in language experience -greater bilingual experience was associated with decreased impact of overlap on memory. We conclude that phonological and semantic features of objects influence memory differently depending on individual differences in language experience, guiding not only what we initially look at, but also what we later remember.

For years, the leading figures of modern AI have been approaching the same phenomenon from different directions.
Some study compression.
Some study neural features.
Some study architectures.
Some study scaling.
At first glance, these appear to be separate research programs. Yet a closer inspection suggests a more intriguing possibility:
They may all be colliding with the same underlying wall.
The wall is not computational.
It is not architectural.
It is not even primarily statistical.
The wall is structural.

This paper presents an assume-guarantee specification theory (aka interface theory from ) for modular synthesis and verification of real-time processes with critical timing constraints. Four operations, i.e. conjunction, disjunction, parallel and quotient, are defined over specifications, drawing inspirations from classic specification theories like refinement calculus . We show that a congruence (or pre-congruence) characterised by a trace-based semantics captures exactly the notion of substitutivity (or refinement) between specifications. Dedication: I would like to thank Prof. Bill Roscoe for leading me into the fascinating world of concurrency and nurturing my appreciation for simplicity and elegance in theories of relevance.

Developmental Evidence for a Canonical Syntax-Semantics Mapping for Verbs of Psychological States Amanda Pogue University of Toronto Joshua K. Hartshorne Harvard University Jesse Snedeker Harvard University Abstract: Psychological state verbs pose problems for theories of the syntax-semantics interface as they can appear with either the stimulus as the subject and the experiencer as the object (SS verbs, e.g. frighten), or vice versa (SO verbs, e.g. fear). In semantic bootstrapping proposals, one of these mapping is treated as an innate default. However theorists disagree about which is the canonical mapping (compare Pinker, 1984 with Grimshaw, 1990) and diary studies suggest the two emerge simultaneously in development (Bowerman, 1990). In Experiment 1, adults, 4- and 5-year-olds showed a trend towards better comprehension of SS-verbs (p=.06), with 4-year-olds performing at chance on SO-verbs. In a less taxing Experiment 2, 4-year-olds again exhibited superior comprehension of SS-v...

This paper surveys the new field of programming methodology and techniques for future quantum computers, including design of sequential and concurrent quantum programming languages, their semantics and implementations. Several verification methods for quantum programs and communication protocols are also reviewed. The potential applications of programming techniques and related formal methods in quantum engineering are pointed out.

This paper argues for a constructionist approach for Aspect by exploring the idea that viewpoint aspect does not exert any altering force on the situation aspect properties of predicates. The proposal is developed by analyzing the point of view where conflicts between situation and viewpoint aspect have been argued to appear in previous literature, namely, the imperfective. The paper focuses on the syntax and semantics of the imperfective, which has been attributed a coercer role as a de-telicizer and de-stativizer in the progressive reading, and as a de-eventizizer in the so-called ability (or attitudinal) and habitual readings. This paper proposes that this is not necessary and provides a unified semantics for the imperfective preserving the properties of eventualities throughout the derivation. The article defends that the semantics of viewpoint aspect is encoded in functional heads containing interval-ordering predicates and quantifiers. This richer structure allows us to analyze aspectual forms with in principle contradictory content such as perfective and progressive, which sheds light onto other issues such as the understanding of non-culminating accomplishments. The proposed syntax is argued to have a corresponding explicit morphology in languages such as Spanish and a non-differentiating one in languages such as English, while the syntax-semantics underlying both of these languages is argued to be the same.

This paper presents a lexical model dedicated to the semantic representation and interpretation of individual words in unrestricted text, where sense discrimination is difficult to assess. We discuss the need of a lexicon including local inference mechanisms and cooperating with as many other knowledge sources (about syntax, semantics and pragmatics) as possible. We suggest a minimalist representation acting as a bridge between a conceptual representation and the microscopic sense variations of lexical semantics. We describe an interpretation method providing one or many alternative candidate(s) to the word, as representatives of its meaning in the sentence (and text). Our system, presently implemented and tested on a restricted scale, belongs to an existing multi-expert architecture for natural language processing, whose ambition is to provide a tool-box for applications involving language understanding.

Semantic technologies and ontologies play an increasing role in scientific workflow systems and knowledge infrastructures. While ontologies are mostly used for the semantic annotation of metadata, semantic technologies enable searching metadata catalogs beyond simple keywords, with some early evidence of semantics used for data translation. However, the next generation of distributed and interdisciplinary knowledge infrastructures will require capabilities beyond simple subsumption reasoning over subclass relations. In this work, we report from the EarthCube Semantics Community by highlighting which role semantics and ontologies should play in the EarthCube knowledge infrastructure. We target the interested domain scientist and, thus, introduce the value proposition of semantic technologies in a non-technical language. Finally, we commit ourselves to some guiding principles for the successful implementation and application of semantic technologies and ontologies within EarthCube.

Modularity has been recognised as a problematic issue of programming language semantics, and various semantic frameworks have been designed with it in mind. Reusability is another desirable feature which, although not the same as modularity, can be enabled by it. The K Framework, based on Rewriting Logic, has good modularity support, but reuse of specifications is not as well developed. The PLanCompS project is developing a framework providing an open-ended collection of reusable components for semantic specification. Each component specifies a single fundamental programming construct, or 'funcon'. The semantics of concrete programming language constructs is given by translating them to combinations of funcons. In this paper, we show how this component-based approach can be seamlessly integrated with the K Framework. We give a component-based definition of CinK (a small subset of C++), using K to define its translation to funcons as well as the (dynamic) semantics of the funcons themselves.

UML solutions in competition on the real-time system market share three common drawbacks: an incomplete formal semantics, temporal operators with limited expression and analysis power, and implementation-oriented tools with limited verification capabilities. To overcome these limitations, the paper proposes a UML profile designed with real-time system validation in mind. Extended class diagrams with associations attributed by composition operators give an explicit semantics to associations between classes. Enhanced activity diagrams with a deterministic delay, a non deterministic delay and a timelimited offering make it possible to work with temporal intervals in lieu of timers with fixed duration. The UML profile is given a precise semantics via its translation into the Formal Description Technique RT-LOTOS. A RT-LOTOS validation tool generates simulation chronograms and reachability graphs for RT-LOTOS specifications derived from UML class and activity diagrams. A coffee machine serves as example. The proposed profile is under evaluation on a satellite-based software reconfiguration system.

This article presents the design, implementation, and experimental validation of an autonomous drone system for search and rescue operations in cluttered GNSS-denied environments. The proposed platform integrates advanced navigation, mapping, and victim-detection capabilities, leveraging a suite of RGB-D cameras and edge-AI computation for real-time perception and decision-making. A key contribution is the integration of an eye-blink-detection pipeline for onboard assessment of the consciousness states of detected victims, enabling the drone to prioritize rescue efforts based on victim alertness. The system employs a modular software architecture with a pipeline that combines a U-Net segmentation network with a MultiScaleLSTM classifier, achieving approximately 97.73% accuracy and a combined inference latency of 6.35 ms on the NVIDIA Jetson Xavier-NX. Experimental results demonstrate the drone’s ability to autonomously explore unknown environments, accurately detect and classify victims, and operate effectively in real-world scenarios. The article also discusses observed challenges, such as computational bottlenecks and false positive detections, and outlines future directions for improving system robustness and autonomy.

An innovative approach for generating Bengali captions from images is introduced in this study, to enhance regional language understanding and vision-language comprehension. Leveraging a substantial Bengali dataset, optimization techniques are incorporated to ensure the efficient implementation of the model. Image features are extracted using established models, including ResNet-50, while captions undergo preprocessing through natural language processing (NLP) techniques. Improved performance is demonstrated by the proposed scheme, which integrates both image and text features, with average parameter values, including BLUE 1: 0.736477, BLUE 2: 0.470831, BLUE 3: 0.252285, BLUE 4: 0.081330, ROUGE: 0.469742, and METEOR: 0.245380. A significant contribution to bridging the gap between regional languages and AI-driven technologies is made by this research, fostering advancements in Bengali vision-language comprehension. The model is designed and experimented with the selection of either InceptionV3, VGG16, or ResNet-50. This work specifically addresses the lack of regionally adapted image captioning systems for Bengali by benchmarking multiple CNN backbones and refining NLP pipelines to tailor performance for a low-resource setting.

This paper describes several results of Wimmics, a research lab which names stands for: web-instrumented man-machine interactions, communities, and semantics. The approaches introduced here rely on graph-oriented knowledge representation, reasoning and operationalization to model and support actors, actions and interactions in web-based epistemic communities. The research results are applied to support and foster interactions in online communities and manage their resources.

The ISICIL initiative (Information Semantic Integration through Communities of Intelligence onLine) mixes viral new web applications with formal semantic web representations and processes to integrate them into corporate practices for technological watch, business intelligence and scientific monitoring. The resulting open source platform proposes three functionalities: (1) a semantic social bookmarking platformmonitored bysemantic social network analysis tools, (2) a system for semantically enriching folksonomies and linking themto corporate terminologies and (3) semantically augmented user interfaces, activity monitoring and reporting tools for business intelligence.

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

We present a new algorithm for the construction of a correlating program from the syntactic difference between the original and modified versions of a program. This correlating program exhibits the semantics of the two input programs and can then be used to compute their semantic differences, following an approach of Partush and Yahav . We show that Partush and Yahav's correlating program is unsound on loops that include an early exit. Our algorithm is defined on an imperative language with while-loops, break, and continue. To guarantee its correctness, it is formalized and mechanically checked within the Coq proof assistant. On a series of examples, we experimentally find that the static analyzer dizy is at least as precise on our correlating program as on Partush and Yahav's.

Object Nets belong to a class of Petri nets allowing for a two-level way of modeling by giving tokens of a Petri net the structure of a Petri net again. The usefulness of this approach has been shown in numerous case studies, ranging from modeling distributed algorithms to workflow and flexible manufacturing systems. It allows for the modeling of real world objects by tokens having their own dynamical behavior. As it is well-known from the field of distributed systems in general, (at least) two different ways of object management are of interest when implementing remote access: either by referencing to a single representation or by creating copies which are treated in a consistent way. In analogy to programming language constructs, this is denoted by reference and value semantics, respectively. In this contribution value and reference semantics of object nets are formally defined. Conditions are presented that allow the transfer from one of these semantics to the other. While the proof techniques strongly rely on partial orders (causal nets), the comparison is made on the basis of occurrence sequences, which serve as a common description language. Objektnetze gehören zu einer Klasse von Petrinetzen, die eine zweistufige Modellierungstechnik dadurch unterstützen, dass die Marken selbst wieder Petrinetze sein dürfen. Dies hat sich bereits in umfangreicher Modellierung für Workflow-und flexible Fertigungs-Systeme als sehr vorteilhaft erwiesen. Objekte der realen Welt wie Aufträge oder Fertigungsteile werden dabei mit ihrem Bearbeitungsplan und ihrer Fortentwicklung dargestellt. Im Kontext verteilter Systeme sind als Zugriffsart für nichtlokale Objekte bekanntlich (mindestens) zwei Formen von Interesse: die Referenz auf eine einzige zentrale Objektrepräsentation oder die Erzeugung und konsistente Pflege von Kopien in unterschiedlichen Systemteilen. In Analogie zur Programmiersprachenterminologie werden diese als Referenzbzw. Wert-Semantik bezeichnet. In diesem Bericht werden Referenzund Wert-Semantik von Objektnetzen formal eingeführt. Außerdem werden Bedingungen für einen konsistenten Wechsel zwischen ihnen bewiesen. Während die Beweise intensiv von partiellen Ordnungen (Kausalnetzen) Gebrauch machen, basiert der Vergleich der Semantiken auf Ausführungsfolgen.

This study investigated the significant challenges presented by Polysemy, or words with multiple related meanings, among first-year Bachelor of Arts in English Language (BAEL) students by profiling the most difficult polysemous words and examining the effectiveness of their meaning selection strategies. Anchored on the Mental Lexicon Theory and Lexical Ambiguity Resolution (LAR) frameworks, the study employed an explanatory sequential mixed-methods design (QUAN → QUAL). Data were collected from 60 1st-year BAEL students using two ambiguity tests, a Likert-scale strategy survey, and open-ended responses. The quantitative results revealed that words with high sense entropy, specifically issue, claim, and record, formed the most persistent difficulty. Although students reported frequent use of prior knowledge and context clues strategies, the findings from the multiple linear regression analysis showed that only contextual strategies significantly predicted accurate meaning selection. Qualitative findings further confirmed that effective polysemous ambiguity resolution depends on integrating contextual constraints with lexical knowledge. These findings highlight the central role of context-based strategies in vocabulary instruction and contribute to pedagogical approaches that strengthen students' lexical processing and reading comprehension skills.

We propose a new directed graphical representation of utility functions, called UCP-networks, that combines aspects of two existing graphical models: generalized additive models and CP-networks. The network decomposes a utility function into a number of additive factors, with the directionality of the arcs reflecting conditional dependence of preference statements - in the underlying (qualitative) preference ordering - under a {em ceteris paribus} (all else being equal) interpretation. This representation is arguably natural in many settings. Furthermore, the strong CP-semantics ensures that computation of optimization and dominance queries is very efficient. We also demonstrate the value of this representation in decision making. Finally, we describe an interactive elicitation procedure that takes advantage of the linear nature of the constraints on "`tradeoff weights" imposed by a UCP-network. This procedure allows the network to be refined until the regret of the decision with minimax regret (with respect to the incompletely specified utility function) falls below a specified threshold (e.g., the cost of further questioning.

Technology is an indispensable part of our day-today activities and major driving force behind revolutionizing every aspect of our lives by providing automated solutions and simplifying various processes. New innovations are no longer limited to providing services and solutions, rather technological advances in IoT and machine learning are making it possible for smart devices collect real time data and translate that data to understand the greater needs of the society, thus being the harbinger of artificial intelligence.

The Problem Frame approach (PF) is an effective requirements engineering tool for analysing and structuring software development problems. PF has a graphical notation that is easy to use and understand. PF imposes only loose constraints on the features of the language used to describe its components and, with only slight restrictions, Hoare's CSP can be used. Although not generally applicable, in this paper we show that CSP can provide a constructive way to arrive at a solution to a PF expressed problem.

Scalable formal analysis of reactive programs demands integration of modular reasoning techniques with existing analysis tools. Principles such as abstraction, compositional refinement, and assume-guarantee reasoning are well understood for architectural hierarchy that describes the communication structure between component processes, and have been shown to be useful. In this paper, we develop the theory of modular reasoning for behavior hierarchy that describes control structure using hierarchic modes. From Statecharts to UML, behavior hierarchy has been an integral component of many software design languages, but only syntactically. We present the hierarchic reactive modules language that retains powerful features such as nested modes, mode reuse, exceptions, group transitions, history, and conjunctive modes, and yet has a semantic notion of mode hierarchy. We present an observational trace semantics for modes that provides the basis for mode refinement. We show the refinement to be compositional with respect to the mode constructors, and develop an assume-guarantee reasoning principle.

Runtime verification (RV) is a pragmatic and scalable, yet rigorous technique, to assess the correctness of complex systems, including cyber-physical systems (CPS). By measuring how robustly a CPS run satisfies a specification, RV allows in addition, to quantify the resiliency of a CPS to perturbations. In this paper we propose Algebraic Runtime Verification (ARV), a general, semantic framework for RV, which takes advantage of the monoidal structure of runs (w.r.t. concatenation) and the semiring structure of a specification automaton (w.r.t. choice and concatenation), to compute in an incremental and application specific fashion the resiliency measure. This allows us to expose the core aspects of RV, by developing an abstract monitoring algorithm, and to strengthen and unify the various qualitative and quantitative approaches to RV, by instantiating choice and concatenation with real-valued functions as dictated by the application. We demonstrate the power and effectiveness of our framework on two case studies from the automotive domain.

Sequential and parallel composition are the most fundamental operators for incremental construction of complex concurrent systems. They reflect the temporal and respectively the spatial properties of these systems. Hiding temporal detail like internal computation steps supports temporal scalability and may turn an asynchronous system to a synchronous one. Hiding spatial detail like internal variables supports spatial scalability and may turn a synchronous system to an asynchronous one. In this paper we show on hand of several examples that a language explicitly supporting both sequential and parallel composition operators is a natural setting for designing heterogeneous synchronous and asynchronous systems. The language we use is Shrm, a visual language that backs up the popular and/or hierarchies of statecharts with a well defined compositional semantics.

In this paper, we demonstrate how a precise semantics of object-oriented modeling techniques can be achieved, and what the possible bene ts are. We outline the main modeling techniques used in the SysLab project, sketch, how a precise semantics can be given, and how this semantics can be used during the development process. ? This paper originates from the SysLab project, which is supported by the DFG under the Leibnizprogramme and by Siemens-Nixdorf.