-
Passage-specific Prompt Tuning for Passage Reranking in Question Answering with Large Language Models
Authors:
Xuyang Wu,
Zhiyuan Peng,
Krishna Sravanthi Rajanala Sai,
Hsin-Tai Wu,
Yi Fang
Abstract:
Effective passage retrieval and reranking methods have been widely utilized to identify suitable candidates in open-domain question answering tasks, recent studies have resorted to LLMs for reranking the retrieved passages by the log-likelihood of the question conditioned on each passage. Although these methods have demonstrated promising results, the performance is notably sensitive to the human-…
▽ More
Effective passage retrieval and reranking methods have been widely utilized to identify suitable candidates in open-domain question answering tasks, recent studies have resorted to LLMs for reranking the retrieved passages by the log-likelihood of the question conditioned on each passage. Although these methods have demonstrated promising results, the performance is notably sensitive to the human-written prompt (or hard prompt), and fine-tuning LLMs can be computationally intensive and time-consuming. Furthermore, this approach limits the leverage of question-passage relevance pairs and passage-specific knowledge to enhance the ranking capabilities of LLMs. In this paper, we propose passage-specific prompt tuning for reranking in open-domain question answering (PSPT): a parameter-efficient method that fine-tunes learnable passage-specific soft prompts, incorporating passage-specific knowledge from a limited set of question-passage relevance pairs. The method involves ranking retrieved passages based on the log-likelihood of the model generating the question conditioned on each passage and the learned soft prompt. We conducted extensive experiments utilizing the Llama-2-chat-7B model across three publicly available open-domain question answering datasets and the results demonstrate the effectiveness of the proposed approach.
△ Less
Submitted 20 June, 2024; v1 submitted 31 May, 2024;
originally announced May 2024.
-
Deep Learning based CNN Model for Classification and Detection of Individuals Wearing Face Mask
Authors:
R. Chinnaiyan,
Iyyappan M,
Al Raiyan Shariff A,
Kondaveeti Sai,
Mallikarjunaiah B M,
P Bharath
Abstract:
In response to the global COVID-19 pandemic, there has been a critical demand for protective measures, with face masks emerging as a primary safeguard. The approach involves a two-fold strategy: first, recognizing the presence of a face by detecting faces, and second, identifying masks on those faces. This project utilizes deep learning to create a model that can detect face masks in real-time str…
▽ More
In response to the global COVID-19 pandemic, there has been a critical demand for protective measures, with face masks emerging as a primary safeguard. The approach involves a two-fold strategy: first, recognizing the presence of a face by detecting faces, and second, identifying masks on those faces. This project utilizes deep learning to create a model that can detect face masks in real-time streaming video as well as images. Face detection, a facet of object detection, finds applications in diverse fields such as security, biometrics, and law enforcement. Various detector systems worldwide have been developed and implemented, with convolutional neural networks chosen for their superior performance accuracy and speed in object detection. Experimental results attest to the model's excellent accuracy on test data. The primary focus of this research is to enhance security, particularly in sensitive areas. The research paper proposes a rapid image pre-processing method with masks centred on faces. Employing feature extraction and Convolutional Neural Network, the system classifies and detects individuals wearing masks. The research unfolds in three stages: image pre-processing, image crop**, and image classification, collectively contributing to the identification of masked faces. Continuous surveillance through webcams or CCTV cameras ensures constant monitoring, triggering a security alert if a person is detected without a mask.
△ Less
Submitted 17 November, 2023;
originally announced November 2023.
-
Information-based Preprocessing of PLC Data for Automatic Behavior Modeling
Authors:
Brandon K. Sai,
Jonas Gram,
Thomas Bauernhansl
Abstract:
Cyber-physical systems (CPS) offer immense optimization potential for manufacturing processes through the availability of multivariate time series data of actors and sensors. Based on automated analysis software, the deployment of adaptive and responsive measures is possible for time series data. Due to the complex and dynamic nature of modern manufacturing, analysis and modeling often cannot be e…
▽ More
Cyber-physical systems (CPS) offer immense optimization potential for manufacturing processes through the availability of multivariate time series data of actors and sensors. Based on automated analysis software, the deployment of adaptive and responsive measures is possible for time series data. Due to the complex and dynamic nature of modern manufacturing, analysis and modeling often cannot be entirely automated. Even machine- or deep learning approaches often depend on a priori expert knowledge and labelling. In this paper, an information-based data preprocessing approach is proposed. By applying statistical methods including variance and correlation analysis, an approximation of the sampling rate in event-based systems and the utilization of spectral analysis, knowledge about the underlying manufacturing processes can be gained prior to modeling. The paper presents, how statistical analysis enables the pruning of a dataset's least important features and how the sampling rate approximation approach sets the base for further data analysis and modeling. The data's underlying periodicity, originating from the cyclic nature of an automated manufacturing process, will be detected by utilizing the fast Fourier transform. This information-based preprocessing method will then be validated for process time series data of cyber-physical systems' programmable logic controllers (PLC).
△ Less
Submitted 15 January, 2024; v1 submitted 28 July, 2023;
originally announced July 2023.
-
A transformer-based deep learning approach for classifying brain metastases into primary organ sites using clinical whole brain MRI
Authors:
Qing Lyu,
Sanjeev V. Namjoshi,
Emory McTyre,
Umit Topaloglu,
Richard Barcus,
Michael D. Chan,
Christina K. Cramer,
Waldemar Debinski,
Metin N. Gurcan,
Glenn J. Lesser,
Hui-Kuan Lin,
Reginald F. Munden,
Boris C. Pasche,
Kiran Kumar Solingapuram Sai,
Roy E. Strowd,
Stephen B. Tatter,
Kounosuke Watabe,
Wei Zhang,
Ge Wang,
Christopher T. Whitlow
Abstract:
Treatment decisions for brain metastatic disease rely on knowledge of the primary organ site, and currently made with biopsy and histology. Here we develop a novel deep learning approach for accurate non-invasive digital histology with whole-brain MRI data. Our IRB-approved single-site retrospective study was comprised of patients (n=1,399) referred for MRI treatment-planning and gamma knife radio…
▽ More
Treatment decisions for brain metastatic disease rely on knowledge of the primary organ site, and currently made with biopsy and histology. Here we develop a novel deep learning approach for accurate non-invasive digital histology with whole-brain MRI data. Our IRB-approved single-site retrospective study was comprised of patients (n=1,399) referred for MRI treatment-planning and gamma knife radiosurgery over 21 years. Contrast-enhanced T1-weighted and T2-weighted Fluid-Attenuated Inversion Recovery brain MRI exams (n=1,582) were preprocessed and input to the proposed deep learning workflow for tumor segmentation, modality transfer, and primary site classification into one of five classes. Ten-fold cross-validation generated overall AUC of 0.878 (95%CI:0.873,0.883), lung class AUC of 0.889 (95%CI:0.883,0.895), breast class AUC of 0.873 (95%CI:0.860,0.886), melanoma class AUC of 0.852 (95%CI:0.842,0.862), renal class AUC of 0.830 (95%CI:0.809,0.851), and other class AUC of 0.822 (95%CI:0.805,0.839). These data establish that whole-brain imaging features are discriminative to allow accurate diagnosis of the primary organ site of malignancy. Our end-to-end deep radiomic approach has great potential for classifying metastatic tumor types from whole-brain MRI images. Further refinement may offer an invaluable clinical tool to expedite primary cancer site identification for precision treatment and improved outcomes.
△ Less
Submitted 20 April, 2022; v1 submitted 7 October, 2021;
originally announced October 2021.
-
Cyclic Shift Problems on Graphs
Authors:
Kwon Kham Sai,
Ryuhei Uehara,
Giovanni Viglietta
Abstract:
We study a new reconfiguration problem inspired by classic mechanical puzzles: a colored token is placed on each vertex of a given graph; we are also given a set of distinguished cycles on the graph. We are tasked with rearranging the tokens from a given initial configuration to a final one by using cyclic shift operations along the distinguished cycles. We first investigate a large class of graph…
▽ More
We study a new reconfiguration problem inspired by classic mechanical puzzles: a colored token is placed on each vertex of a given graph; we are also given a set of distinguished cycles on the graph. We are tasked with rearranging the tokens from a given initial configuration to a final one by using cyclic shift operations along the distinguished cycles. We first investigate a large class of graphs, which generalizes several classic puzzles, and we give a characterization of which final configurations can be reached from a given initial configuration. Our proofs are constructive, and yield efficient methods for shifting tokens to reach the desired configurations. On the other hand, when the goal is to find a shortest sequence of shifting operations, we show that the problem is NP-hard, even for puzzles with tokens of only two different colors.
△ Less
Submitted 23 September, 2020;
originally announced September 2020.
-
Low Cost Autonomous Navigation and Control of a Mechanically Balanced Bicycle with Dual Locomotion Mode
Authors:
Ayush Pandey,
Subhamoy Mahajan,
Adarsh Kosta,
Dhananjay Yadav,
Vikas Pandey,
Saurav Sahay,
Siddharth Jha,
Shubh Agarwal,
Aashay Bhise,
Raushan Kumar,
Aniket Bhushan,
Vraj Parikh,
Ankit Lohani,
Saurabh Dash,
Himanshu Choudhary,
Rahul Kumar,
Anurag Sharma,
Arnab Mondal,
Chendika Karthik Sai,
P N Vamshi
Abstract:
On the lines of the huge and varied efforts in the field of automation with respect to technology development and innovation of vehicles to make them run autonomously, this paper presents an innovation to a bicycle. A normal daily use bicycle was modified at low cost such that it runs autonomously, while maintaining its original form i.e. the manual drive. Hence, a bicycle which could be normally…
▽ More
On the lines of the huge and varied efforts in the field of automation with respect to technology development and innovation of vehicles to make them run autonomously, this paper presents an innovation to a bicycle. A normal daily use bicycle was modified at low cost such that it runs autonomously, while maintaining its original form i.e. the manual drive. Hence, a bicycle which could be normally driven by any human and with a press of switch could run autonomously according to the needs of the user has been developed.
△ Less
Submitted 1 November, 2016;
originally announced November 2016.
