Filters








5,528 Hits in 5.3 sec

Asking without Telling: Exploring Latent Ontologies in Contextual Representations [article]

Julian Michael, Jan A. Botha, Ian Tenney
2020 arXiv   pre-print
Without access to fine-grained gold labels, LSL extracts emergent structure from input representations in an interpretable and quantifiable form.  ...  The success of pretrained contextual encoders, such as ELMo and BERT, has brought a great deal of interest in what these models learn: do they, without explicit supervision, learn to encode meaningful  ...  We treat the classes as clusters representing a latent ontology that describes the underlying representation space.  ... 
arXiv:2004.14513v2 fatcat:4nj2vqn4rnepfdvjnqgny7umai

Asking without Telling: Exploring Latent Ontologies in Contextual Representations

Julian Michael, Jan A. Botha, Ian Tenney
2020 Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)   unpublished
Without access to fine-grained gold labels, LSL extracts emergent structure from input representations in an interpretable and quantifiable form.  ...  The success of pretrained contextual encoders, such as ELMo and BERT, has brought a great deal of interest in what these models learn: do they, without explicit supervision, learn to encode meaningful  ...  We treat the classes as clusters representing a latent ontology that describes the underlying representation space.  ... 
doi:10.18653/v1/2020.emnlp-main.552 fatcat:3v4n7j5navdpxg7fglgpmvwstq

"Tell me more"

Francesco Setti, Daniele Porello, Roberta Ferrario, Sami Abduljalil Abdulhak, Marco Cristani
2013 Proceedings of the International Workshop on Video and Image Ground Truth in Computer Vision Applications - VIGTA '13  
We exploit semantic technologies, such as lexical resources and ontologies, in order to improve the search performances by using a standard web search engine.  ...  In this paper we address the problem of automatically generating a set of images representing an object class, given the name of the class.  ...  In this example, the ontology would tell that the terms "heart" and "stomach" have not to be accepted as keywords for the search, while the term "legs" should be.  ... 
doi:10.1145/2501105.2501110 dblp:conf/vigta/SettiPFAC13 fatcat:bbwykp5ghbbdznxrq4gggwlvpi

PART I: THEORY [chapter]

2018 Telling the Truth  
in the ontolog ical approach to mimesis.  ...  The invention of characters and events possessing differential ontological status thus engages the text in representational procedures that are simultane ously referential and judgmental.  ... 
doi:10.7591/9781501722899-002 fatcat:diosybrn3ze3zb5xtfoi3s4nem

PART II: PRACTICE [chapter]

2018 Telling the Truth  
Robinson Crusoe shows Defoe moving toward undisguised mimetic representation and openly exploring the limits of pseudofactual im posture.  ...  he asked. "By what right do we put the inert first?"  ... 
doi:10.7591/9781501722899-003 fatcat:v65s72i7d5gadfvmpv4ahx5riu

Entangled Biographies: Rebuilding a Sasak House

Kari Telle
2007 Ethnos  
The concept of 'compatibility' has a particular salience in many Indonesian societies. This article examines 'compatibility' with reference to person-house relationships on the island of Lombok.  ...  Terms in brackets are Sasak terms except where I indicate that Indonesian (b.i.) is used. 2. For a discussion of ritual meals (roah), see Telle 2000. roah), see Telle 2000. roah 3.  ...  This article is based on 17 months of fi eldwork conducted between 1993 and 2006 in Jonggat, Central Lombok.  ... 
doi:10.1080/00141840701387903 fatcat:gbm3qklypzb6zkk2ik4z6hbalu

Context-Aware Collaborative Filtering Using Context Similarity: An Empirical Comparison

Yong Zheng
2022 Information  
However, it suffers from the sparsity problem since users may not rate items in various context situations.  ...  A process of context matching, therefore, enables the system to utilize rating profiles in the matched contexts to produce context-aware recommendations.  ...  As mentioned before, the food data is a rating data with dense contextual ratings-each subject was asked for rate selected items in all six contextual situations.  ... 
doi:10.3390/info13010042 fatcat:5lltmj4lobc5bj7s5dou6o4msm

Latent Topology Induction for Understanding Contextualized Representations [article]

Yao Fu, Mirella Lapata
2022 arXiv   pre-print
In this work, we study the representation space of contextualized embeddings and gain insight into the hidden topology of large language models.  ...  We show there exists a network of latent states that summarize linguistic properties of contextualized representations.  ...  Conclusions In this work, we discover a latent state network intrinsically within the representation space of contextualized representations.  ... 
arXiv:2206.01512v1 fatcat:vzefar3urrdt5gyv7mezpcdhve

Exploring Semantically Interlaced Cultural Heritage Narratives

Noemi Mauro, Angelo Geninatti Cossatin, Ester Cravero, Liliana Ardissono, Guido Magnano, Marco Giardino
2022 Proceedings of the 33rd ACM Conference on Hypertext and Social Media  
For this purpose, we propose an exploration model that enables the user to take side walks in semanticallyrelated narrations concerning Points of Interest.  ...  Based on this representation, we developed the Triangolazioni mobile guide that presents multimedia information about Cultural Heritage in Torino city.  ...  Moreover, we asked them to fill in the Curiosity and Exploration Inventory-II (CEI-II) questionnaire [15] . 2) Participants freely used the mobile guide to explore the information about Palazzo Madama  ... 
doi:10.1145/3511095.3536366 fatcat:g5g6htaz2bf7pdrfop4r3zwgi4

Adaptable Conversational Machines

Nurul Lubis, Michael Heck, Carel Van Niekerk, Milica Gasic
2020 The AI Magazine  
In recent years we have witnessed a surge in machine learning methods that provide machines with conversational abilities.  ...  The most prominent mantra in speech and language technology has been "There is no data like more data."  ...  The field has since seen a surge of sophisticated language modeling techniques used to learn latent contextual word and sentence representations.  ... 
doi:10.1609/aimag.v41i3.5322 fatcat:m5grirvy45d7nnqwemcymwq3bq

Ontology-Based Natural Language Query Interfaces for Data Exploration

Chuan Lei, Fatma Özcan, Abdul Quamar, Ashish R. Mittal, Jaydeep Sen, Diptikalyan Saha, Karthik Sankaranarayanan
2018 IEEE Data Engineering Bulletin  
Users, while asking questions to the NLQ interfaces, are latently aware of the context and ask questions keeping that in mind.  ...  A possible way to do this is to use the contextual knowledge from the input query to identify the query is asking about "Loans" and "Loans" does not involve "Person" as per the ontology model.  ... 
dblp:journals/debu/LeiOQMSSS18 fatcat:66g5cfpznnau5kgw3ah4abgtpi

Contextuality, Fine-Tuning, and Teleological Explanation [article]

Emily Adlam
2021 arXiv   pre-print
I assess various proposals for the source of the intuition that there is something problematic about contextuality, ultimately concluding that contextuality is best thought of in terms of fine-tuning.  ...  In the course of this discussion I obtain several new mathematical results - I demonstrate that preparation contextuality is a form of fine-tuning, I show that measurement contextuality can be explained  ...  But the contextuality arises in a very natural way, because measurements in marble world amount to asking systems to 'pick the best option.'  ... 
arXiv:2110.15898v2 fatcat:bqtcn2hnmrhypncjwd6562picu

A Contextualised General Systems Theory

Kirsty Kitto
2014 Systems  
Starting with the notion of measurement, we explore 2 increasingly contextual system behaviour, and identify three major forms of contextuality 3 that might be exhibited by a system: (a) between components  ...  Contextuality must be ontological.  ...  A geometrical 326 account of probability allows for a particularly natural representation of how a system might respond 327 to a change in context (when a state of affairs becomes ontologically different  ... 
doi:10.3390/systems2040541 fatcat:bm34jyz36jawldhdg24tgdkwda

A Survey of Knowledge Enhanced Pre-trained Models [article]

Jian Yang, Gang Xiao, Yulong Shen, Wei Jiang, Xinyu Hu, Ying Zhang, Jinghui Peng
2022 arXiv   pre-print
In this survey, we provide a comprehensive overview of KEPTMs in NLP and CV. We first introduce the progress of pre-trained models and knowledge representation learning.  ...  Pre-trained models learn informative representations on large-scale training data through a self-supervised or supervised learning method, which has achieved promising performance in natural language processing  ...  The reason lies in the learned representations capture both contextual information and knowledge.  ... 
arXiv:2110.00269v3 fatcat:b2g3ezuplvftfp7zlehvogd44m

Ethan Kleinberg: Theory of History as Hauntology

André Da Silva Ramos
2018 History Of Historiography  
The past by definition is gone and thus has no ontological properties or one might say that it has a latent ontology that is activated when we do history.  ...  So one question we need to ask is what holds the ontological certainty of the past event given the possibility of epistemological uncertainty in recounting that event?  ... 
doi:10.15848/hh.v0i25.1263 fatcat:wshckp7oozaada2lj5acccegve
« Previous Showing results 1 — 15 out of 5,528 results