Academic literature on the topic 'Question Decomposition Meaning Representation'

Understanding natural language questions entails the ability to break down a question into the requisite steps for computing its answer. In this work, we introduce a Question Decomposition Meaning Representation (QDMR) for questions. QDMR constitutes the ordered list of steps, expressed through natural language, that are necessary for answering a question. We develop a crowdsourcing pipeline, showing that quality QDMRs can be annotated at scale, and release the Break dataset, containing over 83K pairs of questions and their QDMRs. We demonstrate the utility of QDMR by showing that (a) it can be used to improve open-domain question answering on the HotpotQA dataset, (b) it can be deterministically converted to a pseudo-SQL formal language, which can alleviate annotation in semantic parsing applications. Last, we use Break to train a sequence-to-sequence model with copying that parses questions into QDMR structures, and show that it substantially outperforms several natural baselines.

Much like sentences are composed of words, words themselves are composed of smaller units. For example, the English word questionably can be analyzed as question+ able+ ly. However, this structural decomposition of the word does not directly give us a semantic representation of the word’s meaning. Since morphology obeys the principle of compositionality, the semantics of the word can be systematically derived from the meaning of its parts. In this work, we propose a novel probabilistic model of word formation that captures both the analysis of a word w into its constituent segments and the synthesis of the meaning of w from the meanings of those segments. Our model jointly learns to segment words into morphemes and compose distributional semantic vectors of those morphemes. We experiment with the model on English CELEX data and German DErivBase (Zeller et al., 2013) data. We show that jointly modeling semantics increases both segmentation accuracy and morpheme F1 by between 3% and 5%. Additionally, we investigate different models of vector composition, showing that recurrent neural networks yield an improvement over simple additive models. Finally, we study the degree to which the representations correspond to a linguist’s notion of morphological productivity.

Стаття присвячено обґрунтуванню психологічного механізму розуміння тексту на основі узагаль­нення ґрунтовних сучасних розробок з указаної тематики. Зазначено, що питання про механізм розуміння тексту ще не отримало вичерпної відповіді. Найчастіше для пояснення психологічного механізму розуміння тексту використовують теорію гіпотез, теорію спільності значення, гіпотезу біологічної і функціональної схожості приписування значень, модель контекстуального семантичного обмеження слів, гіпотезу ре­левантності контексту, теорію дзеркальних нейронів та концепцію свідомого смислотворення. Проте використання методів узагальнення, аналізу та синтезу наукової літератури, елементів описового, коре­лятивного та концептуального аналізів дозволило висунути припущення, що психологічним механізмом ро­зуміння тексту є взаємодія сенсу та значення. 
 Науковий аналіз зазначених понять привів до виокремлення їх диференційних ознак та дав підстави узгодити репрезентовані міркування у наступну гіпотезу: в результаті розуміння у свідомості суб’єкта виникають певні організовані уявлення щодо тексту – смисли, які згодом об’єктивуються у вторинний текст. Відтак психологічним механізмом розуміння є процес розкладання значень тексту на їх семантичні компоненти та утворення на цій основі власних вторинних смислів. Критерієм відбору прийнятних смислів може виступати контекст, що підтримує одні смисли та відкидає інші на основі спільності попереднього смислового змісту (Р. Краус, Н. В. Ігнатенко, В. В. Жовтянська, Г. В. Лосик), або ширше – спільний досвід, що формує такий спільний контекст (О. О. Залевська, М. Якобоні). Смисл виникає при зіткненні пропо­нованого текстом значення з наявними у свідомості реципієнта смислами за потреби їх узгодження; значення – загальне, об’єктивне, закріплене за словом, смисл – індивідуальний, суб’єктивний, не закріплений, а лише асоційований, тому він, як правило, ширший за значення.
 The article is devoted to the substantiation of the psychological mechanism of text understanding on the basis of a generalization of the scrupulous modern scientific works. It is noted that the question of the text un­der­standing mechanism has not yet received an exhaustive answer. Often, hypothesis theory, the theory of community significance, the hypothesis of biological and functional similarity of attribution of meanings, the model of contextual semantic restriction of words, the hypothesis of the relevance of the context, the theory of mirror neurons, and the concept of conscious thought-making are used to explain the psychological mechanism of text under­stan­ding. Meanwhile, results of analysis and synthesis of scientific literature, elements of descriptive, correlative and conceptual analysis let propose that the psychological mechanism of text understanding is the interaction of meaning and sense. The scientific analysis of these concepts led to the distinction of their differential features and gave reason to reconcile the representational arguments in the following hypothesis: sense is certain organized ideas about the text, it appears in the mind as a result of text understanding and represents into the secondary text. The psychological mechanism of understanding, respectively, is the process of decomposition of the text meanings into their semantic components and the formation of secondary sense on this basis. A criterion for selecting acceptable sense can be the context that supports some sense and rejects others based on the generality of the prior semantic content (R. Kraus, N. V. Ihnatenko, V. V. Zhovtyans’ka, H. V. Losyk), or more widely – a com­mon expe­rience, which forms such a common context (O. O.Zalevs’ka, M. Yakaboni). The contact the text proposed meaning with sense that existing in the recipient minds for their agreement led to appear the new secondary sense of the of the meanings; meaning is general, objective, fixed by word, sense is individual, subjective, not fixed, but only asso­ciate, therefore, it is usually wider than meaning.

The commonsense question and answering (CSQA) system predicts the right answer based on a comprehensive understanding of the question. Previous research has developed models that use QA pairs, the corresponding evidence, or the knowledge graph as an input. Each method executes QA tasks with representations of pre-trained language models. However, the ability of the pre-trained language model to comprehend completely remains debatable. In this study, adversarial attack experiments were conducted on question-understanding. We examined the restrictions on the question-reasoning process of the pre-trained language model, and then demonstrated the need for models to use the logical structure of abstract meaning representations (AMRs). Additionally, the experimental results demonstrated that the method performed best when the AMR graph was extended with ConceptNet. With this extension, our proposed method outperformed the baseline in diverse commonsense-reasoning QA tasks.

This brief position paper is the central narrative in a theoretic-historical triptych. The first two documents: A (Faceted) Semantic Web? (La Barre, 2011) and Traditions of Facet Theory or a Garden of Forking Paths? (La Barre, forthcoming) amplify a question Brian Vickery posed to the author in 2005. The current proposal, deeply entwined in the same narrative, seeks to continue the agenda of articulation for a primarily North American audience. Instead of description and comparison, a starkly different analytical framework – Walter Benjamin’s notion of ‘aura’ – is deployed to interrogate facet theory along the core dimensions of tradition, uniqueness, and authenticity. The final aspect of this proposal will be an investigation of the aesthetic of redemption that may well serve as the bedrock for operational definitions and functional requirements for facet theory.

A mathematical model of the meaning of the signal is proposed. The meaning is not the signal itself, but its effect on the recipient. Under the action of the signal, the state of the receiver changes, which is the meaning of the signal. The most general mathematical model is the description of the recipient’s state with the help of some mathematical object, and the meaning is modeled by the action of some operator on this object. Various concrete formalisms are considered: abstract automata, matrix representation, algorithms, Markov chains, parameter spaces. The article deals with finite, countable and continuous meanings, reversible and irreversible meanings, ambiguous meanings, decomposition into elementary meanings.

This article intends to critically envisage limits and values of philosophy of art of Arthur Danto and to point out the main problems of the theory of supstitutional representation, when placed within wider theoretical frame of traditional aesthetics, such as the notion of meaning in the philosophy of art of Arthur Danto. The article focuses on the notions of exteral and interal representation and denotation of non-existent and existent entities substituted by representation. This article intends to question the validity of Danto?s position.

Knowledge Graph Question-Answering (KGQA) has gained popularity as an effective approach for information retrieval systems. However, answering complex questions involving multiple topic entities and multi-hop relations presents a significant challenge for model training. Moreover, existing KGQA models face difficulties in extracting constraint information from complex questions, leading to reduced accuracy. To overcome these challenges, we propose a three-part pipelined framework comprising question decomposition, constraint extraction, and question reasoning. Our approach employs a novel question decomposition model that uses dual encoders and attention mechanisms to enhance question representation. We define temporal, spatial, and numerical constraint types and propose a constraint extraction model to mitigate the impact of constraint interference on downstream question reasoning. The question reasoning model uses beam search to reduce computational effort and enhance exploration, facilitating the identification of the optimal path. Experimental results on the ComplexWebQuestions dataset demonstrate the efficacy of our proposed model, achieving an F1 score of 72.0% and highlighting the effectiveness of our approach in decomposing complex questions into simple subsets and improving the accuracy of question reasoning.

In this article, we investigate questions like What is your name again?, which presuppose that the answer was already made common-ground knowledge in the past ( Sauerland 2006 ). We call this a remind-me presupposition. While repetitive particles can trigger a remind-me presupposition in German and English, Japanese uses a specialized particle kke to bring about such a presupposition. We argue for an account of remind-me presuppositions based on syntactic decomposition of the question speech-act into an imperative part and a make-it-known part. On this account, the repetitive particles take scope between the two parts of the decomposed question speech-act. The proposal correctly predicts how both particles interact syntactically with the periphery of the clause in slightly different ways. The interaction with polar questions corroborates our proposal that the decomposed question speech-act parts are syntactically projected parts of the question structure. Our data therefore corroborate a syntactic representation of aspects of speech-acts.

In the article, the authors raise the question of the relationship between the pupils’ perception of the teacher and effective meaning transmission. The study identified socio-psychological qualities and non-verbal patterns of teacher behavior that influence the effectiveness of meaning from the point of view of students. Students prefer to come into contact, with a teacher who has characteristics such as sincerity, leadership, compassion, restraint, responsibility, kindness, cordiality and others. It is established that among the nonverbal patterns of behavior that contribute to the effectiveness of meaning, pupils are considered “gestures appropriate to the situation communication”, “communication distance, which is regulated in accordance with the situation” and “the manifestation of the ability to listen and speak”.