Jose Gallego-Posada, Juan Ramirez de los Rios, Akram Erraqabi
We propose to approach the problem of learning L 0 -sparse networks using a constrained formulation of the optimization problem. This is in contrast to commonly used penalized approaches, which combine the regularization terms additively with the (surrogate) empirical risk. Our experiments demonstrate that we can obtain approximate solutions to the constrained optimization problem with comparable performance to state-of-the art methods for L 0 -sparse training. Finally, we discuss how this constrained approach provides greater (hyper-)parameter interpretability and accountability from a practitioner’s point of view.
{"title":"Flexible Learning of Sparse Neural Networks via Constrained L0 Regularizations","authors":"Jose Gallego-Posada, Juan Ramirez de los Rios, Akram Erraqabi","doi":"10.52591/lxai202112071","DOIUrl":"https://doi.org/10.52591/lxai202112071","url":null,"abstract":"We propose to approach the problem of learning L 0 -sparse networks using a constrained formulation of the optimization problem. This is in contrast to commonly used penalized approaches, which combine the regularization terms additively with the (surrogate) empirical risk. Our experiments demonstrate that we can obtain approximate solutions to the constrained optimization problem with comparable performance to state-of-the art methods for L 0 -sparse training. Finally, we discuss how this constrained approach provides greater (hyper-)parameter interpretability and accountability from a practitioner’s point of view.","PeriodicalId":355096,"journal":{"name":"LatinX in AI at Neural Information Processing Systems Conference 2021","volume":"122 7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124527815","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In modern days, social media platforms provide accessible channels for the inter- 1 action and immediate reflection of the most important events happening around 2 the world. In this paper, we, firstly, present a curated set of datasets whose origin 3 stem from the Twitter’s Information Operations 1 efforts. More notably, these 4 accounts, which have been already suspended, provide a notion of how state-backed 5 human trolls operate. 6 Secondly, we present detailed analyses of how these behaviours vary over time, 7 and motivate its use and abstraction in the context of deep representation learning: 8 for instance, to learn and, potentially track, troll behaviour. We present baselines 9 for such tasks and highlight the differences there may exist within the literature. 10 Finally, we utilize the representations learned for behaviour prediction to classify 11 trolls from "real" users, using a sample of non-suspended active accounts. 12
{"title":"Curating the Twitter Election Integrity Datasets for Better Online Troll Characterization","authors":"Albert Orozco, Reihaneh Rabbany","doi":"10.52591/202112076","DOIUrl":"https://doi.org/10.52591/202112076","url":null,"abstract":"In modern days, social media platforms provide accessible channels for the inter- 1 action and immediate reflection of the most important events happening around 2 the world. In this paper, we, firstly, present a curated set of datasets whose origin 3 stem from the Twitter’s Information Operations 1 efforts. More notably, these 4 accounts, which have been already suspended, provide a notion of how state-backed 5 human trolls operate. 6 Secondly, we present detailed analyses of how these behaviours vary over time, 7 and motivate its use and abstraction in the context of deep representation learning: 8 for instance, to learn and, potentially track, troll behaviour. We present baselines 9 for such tasks and highlight the differences there may exist within the literature. 10 Finally, we utilize the representations learned for behaviour prediction to classify 11 trolls from \"real\" users, using a sample of non-suspended active accounts. 12","PeriodicalId":355096,"journal":{"name":"LatinX in AI at Neural Information Processing Systems Conference 2021","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127059366","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Researchers have extensively used social media platforms like Twitter for knowl-edge discovery purposes, as tweets are considered a wealth of information that provides unique insights. Recent developments have further enabled social media mining for various biomedical tasks such as pharmacovigilance. A first step towards identifying a use-case of Twitter for the pharmacovigilance domain is to extract medication/drug terminologies mentioned in the tweets, which is a challenging task due to several reasons. For example, drug mentions in tweets may be incorrectly written, making the identification of these mentions more difficult. In this work, we propose a two step approach, first, we focused on classifying tweets with drug mentions via an ensemble model (containing transformer models), second, we extract drug mentions (along with their span positions) using a text-tagging/dictionary based approach, and a Named Entity Recognition (NER) approach. By comparing these two entity identification approaches, we demonstrate that using only a dictionary-based approach is not enough.
{"title":"A Pharmacovigilance Application of Social Media Mining: An Ensemble Approach for Automated Classification and Extraction of Drug","authors":"L. Robles, Rajath Chikkatur, J. Banda","doi":"10.52591/202112075","DOIUrl":"https://doi.org/10.52591/202112075","url":null,"abstract":"Researchers have extensively used social media platforms like Twitter for knowl-edge discovery purposes, as tweets are considered a wealth of information that provides unique insights. Recent developments have further enabled social media mining for various biomedical tasks such as pharmacovigilance. A first step towards identifying a use-case of Twitter for the pharmacovigilance domain is to extract medication/drug terminologies mentioned in the tweets, which is a challenging task due to several reasons. For example, drug mentions in tweets may be incorrectly written, making the identification of these mentions more difficult. In this work, we propose a two step approach, first, we focused on classifying tweets with drug mentions via an ensemble model (containing transformer models), second, we extract drug mentions (along with their span positions) using a text-tagging/dictionary based approach, and a Named Entity Recognition (NER) approach. By comparing these two entity identification approaches, we demonstrate that using only a dictionary-based approach is not enough.","PeriodicalId":355096,"journal":{"name":"LatinX in AI at Neural Information Processing Systems Conference 2021","volume":"101 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116519977","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We study the privacy limitations of label differential privacy, which has emerged as an intermediate trust model between local and central differential privacy, where only the label of each training example is protected (and the features are assumed to be public). We show that the guarantees provided by label DP are significantly weaker than they appear, as an adversary can "un-noise" the perturbed labels. Formally we show that the privacy loss has a close connection with Jeffreys’ divergence of the conditional distribution between positive and negative labels, which allows explicit formulation of the trade-off between utility and privacy in this setting. Our results suggest how to select public features that optimize this trade-off. But we still show that there is no free lunch—instances where label differential privacy guarantees are strong are exactly those where a good classifier does not exist. We complement the negative results with a non-parametric estimator for the true privacy loss, and apply our techniques on large-scale benchmark data to demonstrate how to achieve a desired privacy protection.
{"title":"On the Pitfalls of Label Differential Privacy","authors":"Andrés Muñoz","doi":"10.52591/202112077","DOIUrl":"https://doi.org/10.52591/202112077","url":null,"abstract":"We study the privacy limitations of label differential privacy, which has emerged as an intermediate trust model between local and central differential privacy, where only the label of each training example is protected (and the features are assumed to be public). We show that the guarantees provided by label DP are significantly weaker than they appear, as an adversary can \"un-noise\" the perturbed labels. Formally we show that the privacy loss has a close connection with Jeffreys’ divergence of the conditional distribution between positive and negative labels, which allows explicit formulation of the trade-off between utility and privacy in this setting. Our results suggest how to select public features that optimize this trade-off. But we still show that there is no free lunch—instances where label differential privacy guarantees are strong are exactly those where a good classifier does not exist. We complement the negative results with a non-parametric estimator for the true privacy loss, and apply our techniques on large-scale benchmark data to demonstrate how to achieve a desired privacy protection.","PeriodicalId":355096,"journal":{"name":"LatinX in AI at Neural Information Processing Systems Conference 2021","volume":"335 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115395813","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hector Mejia, E. Palomo, Ezequiel López-Rubio, Israel Pineda, R. Fonseca
Currently, the standard for vehicle speed estimation is radar or lidar speed signs which can be costly to buy and maintain. However, most major cities already implement networks of traffic surveillance cameras that can be utilized for vehicle speed estimation using computer vision. This work implements such a system using homography estimation, YOLOv4 object detector, and an object tracker capable of vehicle speed estimation. The homography component uses world plane-image plane point correspondences, located by humans. Moreover, a new method is developed specifically for this use case, using the estimation of density evolutionary algorithm. It aims at correcting the points misalignment in between planes. In addition, a basic direct linear transformation (DLT) and a random sample consensus robust version of DLT are implemented for comparison. Finally, the results show that the proposed homography method reduces the projection error from world to image point by 97%, when compared to the other two methods, and the complete workflow can successfully estimate speed distributions expected from vehicles on urban traffic and handle dynamic changes in vehicle speed.
{"title":"Vehicle Speed Estimation Using Computer Vision and Evolutionary Camera Calibration","authors":"Hector Mejia, E. Palomo, Ezequiel López-Rubio, Israel Pineda, R. Fonseca","doi":"10.52591/lxai202112072","DOIUrl":"https://doi.org/10.52591/lxai202112072","url":null,"abstract":"Currently, the standard for vehicle speed estimation is radar or lidar speed signs which can be costly to buy and maintain. However, most major cities already implement networks of traffic surveillance cameras that can be utilized for vehicle speed estimation using computer vision. This work implements such a system using homography estimation, YOLOv4 object detector, and an object tracker capable of vehicle speed estimation. The homography component uses world plane-image plane point correspondences, located by humans. Moreover, a new method is developed specifically for this use case, using the estimation of density evolutionary algorithm. It aims at correcting the points misalignment in between planes. In addition, a basic direct linear transformation (DLT) and a random sample consensus robust version of DLT are implemented for comparison. Finally, the results show that the proposed homography method reduces the projection error from world to image point by 97%, when compared to the other two methods, and the complete workflow can successfully estimate speed distributions expected from vehicles on urban traffic and handle dynamic changes in vehicle speed.","PeriodicalId":355096,"journal":{"name":"LatinX in AI at Neural Information Processing Systems Conference 2021","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114686002","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：