Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

"Topic": models, code, and papers

Pronoun Resolution in Japanese Sentences Using Surface Expressions and Examples

Dec 13, 1999
M. Murata, H. Isahara, M. Nagao

In this paper, we present a method of estimating referents of demonstrative pronouns, personal pronouns, and zero pronouns in Japanese sentences using examples, surface expressions, topics and foci. Unlike conventional work which was semantic markers for semantic constraints, we used examples for semantic constraints and showed in our experiments that examples are as useful as semantic markers. We also propose many new methods for estimating referents of pronouns. For example, we use the form ``X of Y'' for estimating referents of demonstrative adjectives. In addition to our new methods, we used many conventional methods. As a result, experiments using these methods obtained a precision rate of 87% in estimating referents of demonstrative pronouns, personal pronouns, and zero pronouns for training sentences, and obtained a precision rate of 78% for test sentences.

* ACL'99 Workshop on 'Coreference and Its Applications', Maryland, USA, June 22, 1999 
* 8 pages, 0 figures. Computation and Language 

  Access Paper or Ask Questions

MM-Claims: A Dataset for Multimodal Claim Detection in Social Media

May 04, 2022
Gullal S. Cheema, Sherzod Hakimov, Abdul Sittar, Eric Müller-Budack, Christian Otto, Ralph Ewerth

In recent years, the problem of misinformation on the web has become widespread across languages, countries, and various social media platforms. Although there has been much work on automated fake news detection, the role of images and their variety are not well explored. In this paper, we investigate the roles of image and text at an earlier stage of the fake news detection pipeline, called claim detection. For this purpose, we introduce a novel dataset, MM-Claims, which consists of tweets and corresponding images over three topics: COVID-19, Climate Change and broadly Technology. The dataset contains roughly 86000 tweets, out of which 3400 are labeled manually by multiple annotators for the training and evaluation of multimodal models. We describe the dataset in detail, evaluate strong unimodal and multimodal baselines, and analyze the potential and drawbacks of current models.

* Accepted to Findings of NAACL 2022 

  Access Paper or Ask Questions

CTMSTOU driven markets: simulated environment for regime-awareness in trading policies

Feb 03, 2022
Selim Amrouni, Aymeric Moulin, Tucker Balch

Market regimes is a popular topic in quantitative finance even though there is little consensus on the details of how they should be defined. They arise as a feature both in financial market prediction problems and financial market task performing problems. In this work we use discrete event time multi-agent market simulation to freely experiment in a reproducible and understandable environment where regimes can be explicitly switched and enforced. We introduce a novel stochastic process to model the fundamental value perceived by market participants: Continuous-Time Markov Switching Trending Ornstein-Uhlenbeck (CTMSTOU), which facilitates the study of trading policies in regime switching markets. We define the notion of regime-awareness for a trading agent as well and illustrate its importance through the study of different order placement strategies in the context of order execution problems.

* fix typo in title 

  Access Paper or Ask Questions

A multi-domain virtual network embedding algorithm with delay prediction

Feb 03, 2022
Peiying Zhang, Xue Pang, Yongjing Ni, Haipeng Yao, Xin Li

Virtual network embedding (VNE) is an crucial part of network virtualization (NV), which aims to map the virtual networks (VNs) to a shared substrate network (SN). With the emergence of various delay-sensitive applications, how to improve the delay performance of the system has become a hot topic in academic circles. Based on extensive research, we proposed a multi-domain virtual network embedding algorithm based on delay prediction (DP-VNE). Firstly, the candidate physical nodes are selected by estimating the delay of virtual requests, then particle swarm optimization (PSO) algorithm is used to optimize the mapping process, so as to reduce the delay of the system. The simulation results show that compared with the other three advanced algorithms, the proposed algorithm can significantly reduce the system delay while keeping other indicators unaffected.

  Access Paper or Ask Questions

Document AI: Benchmarks, Models and Applications

Nov 16, 2021
Lei Cui, Yiheng Xu, Tengchao Lv, Furu Wei

Document AI, or Document Intelligence, is a relatively new research topic that refers to the techniques for automatically reading, understanding, and analyzing business documents. It is an important research direction for natural language processing and computer vision. In recent years, the popularity of deep learning technology has greatly advanced the development of Document AI, such as document layout analysis, visual information extraction, document visual question answering, document image classification, etc. This paper briefly reviews some of the representative models, tasks, and benchmark datasets. Furthermore, we also introduce early-stage heuristic rule-based document analysis, statistical machine learning algorithms, and deep learning approaches especially pre-training methods. Finally, we look into future directions for Document AI research.

  Access Paper or Ask Questions

Design of Tight Minimum-Sidelobe Windows by Riemannian Newton's Method

Nov 02, 2021
Daichi Kitahara, Kohei Yatabe

The short-time Fourier transform (STFT), or the discrete Gabor transform (DGT), has been extensively used in signal analysis and processing. Their properties are characterized by a window function, and hence window design is a significant topic up to date. For signal processing, designing a pair of analysis and synthesis windows is important because results of processing in the time-frequency domain are affected by both of them. A tight window is a special window that can perfectly reconstruct a signal by using it for both analysis and synthesis. It is known to make time-frequency-domain processing robust to error, and therefore designing a better tight window is desired. In this paper, we propose a method of designing tight windows that minimize the sidelobe energy. It is formulated as an optimization problem on an oblique manifold, and a Riemannian Newton algorithm on this manifold is derived to efficiently obtain a solution.

  Access Paper or Ask Questions

LAE : Long-tailed Age Estimation

Oct 25, 2021
Zenghao Bao, Zichang Tan, Yu Zhu, Jun Wan, Xibo Ma, Zhen Lei, Guodong Guo

Facial age estimation is an important yet very challenging problem in computer vision. To improve the performance of facial age estimation, we first formulate a simple standard baseline and build a much strong one by collecting the tricks in pre-training, data augmentation, model architecture, and so on. Compared with the standard baseline, the proposed one significantly decreases the estimation errors. Moreover, long-tailed recognition has been an important topic in facial age datasets, where the samples often lack on the elderly and children. To train a balanced age estimator, we propose a two-stage training method named Long-tailed Age Estimation (LAE), which decouples the learning procedure into representation learning and classification. The effectiveness of our approach has been demonstrated on the dataset provided by organizers of Guess The Age Contest 2021.

* The 1st Place in Guess The Age Contest, CAIP2021 (The 19th International Conference on Computer Analysis of Images and Patterns) 

  Access Paper or Ask Questions

Learning Ideological Embeddings from Information Cascades

Sep 28, 2021
Corrado Monti, Giuseppe Manco, Cigdem Aslay, Francesco Bonchi

Modeling information cascades in a social network through the lenses of the ideological leaning of its users can help understanding phenomena such as misinformation propagation and confirmation bias, and devising techniques for mitigating their toxic effects. In this paper we propose a stochastic model to learn the ideological leaning of each user in a multidimensional ideological space, by analyzing the way politically salient content propagates. In particular, our model assumes that information propagates from one user to another if both users are interested in the topic and ideologically aligned with each other. To infer the parameters of our model, we devise a gradient-based optimization procedure maximizing the likelihood of an observed set of information cascades. Our experiments on real-world political discussions on Twitter and Reddit confirm that our model is able to learn the political stance of the social media users in a multidimensional ideological space.

* Proceedings of the 30th ACM International Conference on Information and Knowledge Management (CIKM 2021) 
* Published in CIKM 2021 

  Access Paper or Ask Questions

Not Quite 'Ask a Librarian': AI on the Nature, Value, and Future of LIS

Jul 07, 2021
Jesse David Dinneen, Helen Bubinger

AI language models trained on Web data generate prose that reflects human knowledge and public sentiments, but can also contain novel insights and predictions. We asked the world's best language model, GPT-3, fifteen difficult questions about the nature, value, and future of library and information science (LIS), topics that receive perennial attention from LIS scholars. We present highlights from its 45 different responses, which range from platitudes and caricatures to interesting perspectives and worrisome visions of the future, thus providing an LIS-tailored demonstration of the current performance of AI language models. We also reflect on the viability of using AI to forecast or generate research ideas in this way today. Finally, we have shared the full response log online for readers to consider and evaluate for themselves.

* Final version to appear in ASIS&T '21: Proceedings of the 84th Annual Meeting of the Association for Information Science & Technology, 58 

  Access Paper or Ask Questions

Intent Features for Rich Natural Language Understanding

Apr 21, 2021
Brian Lester, Sagnik Ray Choudhury, Rashmi Prasad, Srinivas Bangalore

Complex natural language understanding modules in dialog systems have a richer understanding of user utterances, and thus are critical in providing a better user experience. However, these models are often created from scratch, for specific clients and use cases, and require the annotation of large datasets. This encourages the sharing of annotated data across multiple clients. To facilitate this we introduce the idea of intent features: domain and topic agnostic properties of intents that can be learned from the syntactic cues only, and hence can be shared. We introduce a new neural network architecture, the Global-Local model, that shows significant improvement over strong baselines for identifying these features in a deployed, multi-intent natural language understanding module, and, more generally, in a classification setting where a part of an utterance has to be classified utilizing the whole context.

* Camera-ready for NAACL 2021 Industry Track 

  Access Paper or Ask Questions