Patent Automatic Classification Based on Symmetric Hierarchical Convolution Neural Network

Zhu, Huiming; He, Chunhui; Fang, Yang; Ge, Bin; Xing, Meng; Xiao, Weidong

doi:10.3390/sym12020186

Cited by 12 publications

(7 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…RNNs can predict time series and have been used successfully for text normalization, de-identification, and sequence labeling [ 61 , 62 ]. However, RNNs also suffer from low efficiency in multi-classifier settings [ 63 ]. In this report, we proposed a novel network design to extract drug compounds from sentences using a regression loss function (i.e., mean square root error (MSE)).…”

Section: Discussionmentioning

confidence: 99%

Adera2.0: A Drug Repurposing Workflow for Neuroimmunological Investigations Using Neural Networks

et al. 2022

View full text Add to dashboard Cite

Drug repurposing in the context of neuroimmunological (NI) investigations is still in its primary stages. Drug repurposing is an important method that bypasses lengthy drug discovery procedures and focuses on discovering new usages for known medications. Neuroimmunological diseases, such as Alzheimer’s, Parkinson’s, multiple sclerosis, and depression, include various pathologies that result from the interaction between the central nervous system and the immune system. However, the repurposing of NI medications is hindered by the vast amount of information that needs mining. We previously presented Adera1.0, which was capable of text mining PubMed for answering query-based questions. However, Adera1.0 was not able to automatically identify chemical compounds within relevant sentences. To challenge the need for repurposing known medications for neuroimmunological diseases, we built a deep neural network named Adera2.0 to perform drug repurposing. The workflow uses three deep learning networks. The first network is an encoder and its main task is to embed text into matrices. The second network uses a mean squared error (MSE) loss function to predict answers in the form of embedded matrices. The third network, which constitutes the main novelty in our updated workflow, also uses a MSE loss function. Its main usage is to extract compound names from relevant sentences resulting from the previous network. To optimize the network function, we compared eight different designs. We found that a deep neural network consisting of an RNN neural network and a leaky ReLU could achieve 0.0001 loss and 67% sensitivity. Additionally, we validated Adera2.0’s ability to predict NI drug usage against the DRUG Repurposing Hub database. These results establish the ability of Adera2.0 to repurpose drug candidates that can shorten the development of the drug cycle. The workflow could be download online.

show abstract

Section: Discussionmentioning

confidence: 99%

Adera2.0: A Drug Repurposing Workflow for Neuroimmunological Investigations Using Neural Networks

et al. 2022

View full text Add to dashboard Cite

show abstract

“…LSTM units realistically represent or simulate the cognitive processes of human behavior, logical development, and neural organization. It is suitable for handling problems that are highly related to time series, such as machine translation, dialogue generation, encoding, and decoding [ 18 ]. The LSTM unit structure is shown in Figure 2 .…”

Section: Methodsmentioning

confidence: 99%

Predictive Control of the Mobile Robot under the Deep Long-Short Term Memory Neural Network Model

Zheng

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

At present, there is a phenomenon of network data packet loss in the trajectory tracking control system, which will degrade or even destabilize the system’s performance. Therefore, this work first explains the theory of the deep long-short term memory (LSTM) neural network model, the kinematic model of mobile robots, and the trajectory tracking error model. The reasons for data packet loss in the control system are analyzed. Second, a prediction model based on the LSTM network is designed according to the theory mentioned above. Finally, the training effect of the LSTM model and the robot trajectory tracking effect based on the model are tested by setting up simulation experiments. The research results are as follows: (1) The pose test error of the mobile robot will eventually tend to zero through the simulation curve generated by the pose parameters (x, y, θ) of the mobile robot. (2) The trajectory tracking error of the deep LSTM neural network prediction and compensation method with the packet loss rate of 5% is less than that with the packet loss rate of 10%. (3) The linear velocity υ of the mobile robot based on the prediction model of the LSTM network varies greatly but is always in the interval (−2, 2). Its angular velocity ω initially fluctuates greatly but gradually tends to zero after about 13 s. (4) When the prediction model tracks the trajectory of the robot, the horizontal position x, the vertical position y, and the angle θ coincide with the reference trajectory. The exploration is conducted to provide a reference for the research on data packet loss in the networked mobile robot trajectory tracking system.

show abstract

“…DeepPatent performs multi-label classification at the sub-class level getting 75.46% recall at top 4. Likewise, Zhu et al [7] used the Word Embedding technique to segment and vectorize the input data and then a symmetric hierarchical CNN, named PAC-HCNN, to classify patents outperforming traditional RNN. Moreover, Abdelgawad et al [8] compared several recent neural network models and showed that CNNs are a suitable choice for patent classification.…”

Section: Prior Workmentioning

confidence: 99%

“…Last but not least, some researchers tried to identify which parts of a patent document can provide more representative information for classification tasks [15,17]. They showed that the use of technical and background parts extracted from the descriptions section [17], the first claim [15], the title and abstract [6,7], or at most cases, the title, abstract claims and description [8,11,12] are the most useful parts for patent classification.…”

Section: Prior Workmentioning

confidence: 99%

“…Recently, automated patent classification, which falls under the broad field of text classification [4,5], has been examined using deep learning (DL) methods such as convolutional neural networks (CNN) [6][7][8], Word2Vec and Long-short term memory (LSTM) [9][10][11] and other DL methods [12] to predict the most representative classification code(s) for a patent. In the current work, these CNN and RNN models are evaluated employing different language models and different parts of the patent document.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Automated Single-Label Patent Classification using Ensemble Classifiers

Kamateri

Stamatis

Diamantaras

et al. 2022

2022 14th International Conference on Machine Learning and Computing (ICMLC)

View full text Add to dashboard Cite

Many thousands of patent applications arrive at patent offices around the world every day. One important task when a patent application is submitted is to assign one or more classification codes from the complex and hierarchical patent classification schemes that will enable routing of the patent application to a patent examiner who is knowledgeable about the specific technical field. This task is typically undertaken by patent professionals, however due to the large number of applications and the potential complexity of an invention, they are usually overwhelmed. Therefore, there is a need for this code assignment manual task to be supported or even fully automated by classification systems that will classify patent applications, hopefully with an accuracy close to patent professionals. Like in many other text analysis problems, in the last years, this intellectually demanding task has been studied using word embeddings and deep learning techniques. In this paper these research efforts are shortly reviewed and re-produced with similar deep learning techniques using different feature representations on automatic patent classification in the level of sub-classes. On top of that, an innovative method of ensemble classifiers trained with different parts of the patent document is proposed. To the best of our knowledge, this is the first time that an ensemble method was proposed for the patent classification problem. Our first results are quite promising showing that an ensemble architecture of classifiers significantly outperforms current state-of-the-art techniques using the same classifiers as standalone solutions.

show abstract

Patent Automatic Classification Based on Symmetric Hierarchical Convolution Neural Network

Cited by 12 publications

References 15 publications

Adera2.0: A Drug Repurposing Workflow for Neuroimmunological Investigations Using Neural Networks

Adera2.0: A Drug Repurposing Workflow for Neuroimmunological Investigations Using Neural Networks

Predictive Control of the Mobile Robot under the Deep Long-Short Term Memory Neural Network Model

Automated Single-Label Patent Classification using Ensemble Classifiers

Contact Info

Product

Resources

About