Task-oriented Domain-specific Meta-Embedding for Text Classification

Today’s world is mostly data-driven. To deal with the humongous amount of data, Machine Learning and Data Mining strategies are put into usage. Traditional ML approaches presume that the model is tested on a dataset extracted from the same domain from where the training data has been taken from. Nevertheless, some real-world situations require machines to provide good results with very little domain-specific training data. This creates room for the development of machines that are capable of predicting accurately by being trained on easily found data. Transfer Learning is the key to it. It is the scientific art of applying the knowledge gained while learning a task to another task that is similar to the previous one in some or another way. This article focuses on building a model that is capable of differentiating text data into binary classes; one roofing the text data that is spam and the other not containing spam using BERT’s pre-trained model (bert-base-uncased). This pre-trained model has been trained on Wikipedia and Book Corpus data and the goal of this paper is to highlight the pre-trained model’s capabilities to transfer the knowledge that it has learned from its training (Wiki and Book Corpus) to classifying spam texts from the rest.

Download Full-text

Domain-Specific Multi-Agent Dialog Policy Learning in Multi-Domain Task-Oriented Scenarios

10.21437/interspeech.2021-887 ◽

2021 ◽

Author(s):

Li Tang ◽

Yuke Si ◽

Longbiao Wang ◽

Jianwu Dang

Keyword(s):

Policy Learning ◽

Domain Specific ◽

Multi Agent ◽

Task Oriented

Download Full-text

An automated domain specific stop word generation method for natural language text classification

2011 International Symposium on Innovations in Intelligent Systems and Applications ◽

10.1109/inista.2011.5946149 ◽

2011 ◽

Cited By ~ 6

Author(s):

H. Ayral ◽

S. Yavuz

Keyword(s):

Natural Language ◽

Text Classification ◽

Word Generation ◽

Natural Language Text ◽

Domain Specific ◽

Stop Word ◽

Language Text

Download Full-text

Text classification based filters for a domain-specific search engine

Computers in Industry ◽

10.1016/j.compind.2015.10.004 ◽

2016 ◽

Vol 78 ◽

pp. 70-79 ◽

Cited By ~ 6

Author(s):

Sebastian Schmidt ◽

Steffen Schnitzer ◽

Christoph Rensing

Keyword(s):

Search Engine ◽

Text Classification ◽

Domain Specific ◽

Specific Search

Download Full-text

Task-Based Learning via Task-Oriented Prediction Network with Applications in Finance

Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2020/617 ◽

2020 ◽

Author(s):

Di Chen ◽

Yada Zhu ◽

Xiaodong Cui ◽

Carla Gomes

Keyword(s):

Decision Making ◽

Real World ◽

Evaluation Criteria ◽

Risk Modeling ◽

Domain Specific ◽

Learning Scheme ◽

Gradient Based ◽

End To End ◽

Task Oriented ◽

And Task

Real-world applications often involve domain-specific and task-based performance objectives that are not captured by the standard machine learning losses, but are critical for decision making. A key challenge for direct integration of more meaningful domain and task-based evaluation criteria into an end-to-end gradient-based training process is the fact that often such performance objectives are not necessarily differentiable and may even require additional decision-making optimization processing. We propose the Task-Oriented Prediction Network (TOPNet), an end-to-end learning scheme that automatically integrates task-based evaluation criteria into the learning process via a learnable surrogate loss function, which directly guides the model towards the task-based goal. A major benefit of the proposed TOPNet learning scheme lies in its capability of automatically integrating non-differentiable evaluation criteria, which makes it particularly suitable for diversified and customized task-based evaluation criteria in real-world tasks. We validate the performance of TOPNet on two real-world financial prediction tasks, revenue surprise forecasting and credit risk modeling. The experimental results demonstrate that TOPNet significantly outperforms both traditional modeling with standard losses and modeling with hand-crafted heuristic differentiable surrogate losses.

Download Full-text

Interacting orientations and instrumentalities to adapt a learning tool for health professionals

Knowledge Management & E-Learning: An International Journal ◽

10.34105/j.kmel.2015.07.032 ◽

2015 ◽

pp. 489-506

Keyword(s):

Knowledge Domain ◽

Web Based ◽

Domain Specific ◽

Service Training ◽

Dialectical Relationship ◽

Cultural Historical Activity ◽

Adaptation Processes ◽

Task Oriented ◽

Relationship Of ◽

Cultural Tools

Web-based instructional software offers new opportunities for collaborative, task-oriented in-service training. Planning and negotiation of content to adapt a web-based learning resource for nursing is the topic of this paper. We draw from Cultural Historical Activity Theory to elaborate the dialectical relationship of changing and stabilizing organizational practice. Local adaptation to create a domain-specific resource plays out as interactions of orientations and instrumentalities. Our analysis traces how orientations, i.e., in situ selection of knowledge and mobilization of experiences, and instrumentality, i.e., interpreted affordances of available cultural tools, interact. The adaptation processes are mediated by a set of new and current tools that interact with multiple orientations to ensure stability and promote change. Practice and project are introduced as intermediate, analytic concepts to assess tensions in the observed activity. Our analysis shows three central tensions, how they are introduced, addressed and subsequently resolved. Considering the opportunities help understand how engagement with technology can lead to new representations for introduction to a local knowledge domain.

Download Full-text