-
DetIE: Multilingual Open Information Extraction Inspired by Object Detection
Authors:
Michael Vasilkovsky,
Anton Alekseev,
Valentin Malykh,
Ilya Shenbin,
Elena Tutubalina,
Dmitriy Salikhov,
Mikhail Stepnov,
Andrey Chertok,
Sergey Nikolenko
Abstract:
State of the art neural methods for open information extraction (OpenIE) usually extract triplets (or tuples) iteratively in an autoregressive or predicate-based manner in order not to produce duplicates. In this work, we propose a different approach to the problem that can be equally or more successful. Namely, we present a novel single-pass method for OpenIE inspired by object detection algorith…
▽ More
State of the art neural methods for open information extraction (OpenIE) usually extract triplets (or tuples) iteratively in an autoregressive or predicate-based manner in order not to produce duplicates. In this work, we propose a different approach to the problem that can be equally or more successful. Namely, we present a novel single-pass method for OpenIE inspired by object detection algorithms from computer vision. We use an order-agnostic loss based on bipartite matching that forces unique predictions and a Transformer-based encoder-only architecture for sequence labeling. The proposed approach is faster and shows superior or similar performance in comparison with state of the art models on standard benchmarks in terms of both quality metrics and inference time. Our model sets the new state of the art performance of 67.7% F1 on CaRB evaluated as OIE2016 while being 3.35x faster at inference than previous state of the art. We also evaluate the multilingual version of our model in the zero-shot setting for two languages and introduce a strategy for generating synthetic multilingual data to fine-tune the model for each specific language. In this setting, we show performance improvement 15% on multilingual Re-OIE2016, reaching 75% F1 for both Portuguese and Spanish languages. Code and models are available at https://github.com/sberbank-ai/DetIE.
△ Less
Submitted 24 June, 2022;
originally announced June 2022.
-
An empirical analysis of success factors in the adaption of the scaled agile framework -- first outcomes from an empirical study
Authors:
Dilshat Salikhov,
Giancarlo Succi,
Alexander Tormasov
Abstract:
Agile methodologies are used for improving productivity and quality of development originally created for small teams. However , now they are expanding to larger organizations, for which "scaled up" approaches have been proposed. This study presents the preliminary outcomes from a survey on the effects of the Scaled Agile Framework (SAFe), which is considered the most used of such approaches. The…
▽ More
Agile methodologies are used for improving productivity and quality of development originally created for small teams. However , now they are expanding to larger organizations, for which "scaled up" approaches have been proposed. This study presents the preliminary outcomes from a survey on the effects of the Scaled Agile Framework (SAFe), which is considered the most used of such approaches. The apparent advantages and limitations are discussed along with the lines for future research.
△ Less
Submitted 21 December, 2020;
originally announced December 2020.
-
Jolie Good Buildings: Internet of things for smart building infrastructure supporting concurrent apps utilizing distributed microservices
Authors:
Kamill Gusmanov,
Kevin Khanda,
Dilshat Salikhov,
Manuel Mazzara,
Nikolaos Mavridis
Abstract:
A large percentage of buildings, domestic or special-purpose, is expected to become increasingly "smarter" in the future, due to the immense benefits in terms of energy saving, safety, flexibility, and comfort, that relevant new technologies offer. However, concerning the hardware, software, or platform levels, no clearly dominant standard frameworks currently exist. Here, we will present a protot…
▽ More
A large percentage of buildings, domestic or special-purpose, is expected to become increasingly "smarter" in the future, due to the immense benefits in terms of energy saving, safety, flexibility, and comfort, that relevant new technologies offer. However, concerning the hardware, software, or platform levels, no clearly dominant standard frameworks currently exist. Here, we will present a prototype platform for supporting multiple concurrent applications for smart buildings, which is utilizing an advanced sensor network as well as a distributed micro services architecture, centrally featuring the Jolie language. The architecture and benefits of our system are discussed, as well as a prototype containing a number of nodes and a user interface, deployed in a real-world academic building environment. Our results illustrate the promising nature of our approach, as well as open avenues for future work towards it wider and larger scale applicability.
△ Less
Submitted 28 November, 2016;
originally announced November 2016.
-
Microservice-based IoT for Smart Buildings
Authors:
Dilshat Salikhov,
Kevin Khanda,
Kamill Gusmanov,
Manuel Mazzara,
Nikolaos Mavridis
Abstract:
A large percentage of buildings in domestic or special-purpose is expected to become increasingly "smarter" in the future, due to the immense benefits in terms of energy saving, safety, flexibility, and comfort, that relevant new technologies offer. As concerns hardware, software, or platform level, however, no clearly dominant standards currently exist. Such standards, would ideally, fulfill a nu…
▽ More
A large percentage of buildings in domestic or special-purpose is expected to become increasingly "smarter" in the future, due to the immense benefits in terms of energy saving, safety, flexibility, and comfort, that relevant new technologies offer. As concerns hardware, software, or platform level, however, no clearly dominant standards currently exist. Such standards, would ideally, fulfill a number of important desiderata, which are to be touched upon in this paper. Here, we will present a prototype platform for supporting multiple concurrent applications for smart buildings, which is utilizing an advanced sensor network as well as a distributed microservices architecture, centrally featuring the Jolie programming language. The architecture and benefits of our system are discussed, as well as a prototype containing a number of nodes and a user interface, deployed in a real-world academic building environment. Our results illustrate the promising nature of our approach, as well as open avenues for future work towards its wider and larger scale applicability.
△ Less
Submitted 29 October, 2016;
originally announced October 2016.
