Machine learning models are bounded by the credibility of ground truth data used for both training and testing. Regardless of the problem domain, this ground truth annotation is objectively manual and tedious as it needs considerable amount of human intervention. With the advent of Active Learning with multiple annotators, the burden can be somewhat mitigated by actively acquiring labels of most informative data instances. However, multiple annotators with varying degrees of expertise poses new set of challenges in terms of quality of the label received and availability of the annotator. Due to limited amount of ground truth information addressing the variabilities of Activity of Daily Living (ADLs), activity recognition models using wearable and mobile devices are still not robust enough for real-world deployment. In this paper, we first propose an active learning combined deep model which updates its network parameters based on the optimization of a joint loss function. We then propose a novel annotator selection model by exploiting the relationships among the users while considering their heterogeneity with respect to their expertise, physical and spatial context. Our proposed model leverages model-free deep reinforcement learning in a partially observable environment setting to capture the action-reward interaction among multiple annotators. Our experiments in real-world settings exhibit that our active deep model converges to optimal accuracy with fewer labeled instances and achieves ~8% improvement in accuracy in fewer iterations.
{"title":"Active Deep Learning for Activity Recognition with Context Aware Annotator Selection","authors":"H. S. Hossain, Nirmalya Roy","doi":"10.1145/3292500.3330688","DOIUrl":"https://doi.org/10.1145/3292500.3330688","url":null,"abstract":"Machine learning models are bounded by the credibility of ground truth data used for both training and testing. Regardless of the problem domain, this ground truth annotation is objectively manual and tedious as it needs considerable amount of human intervention. With the advent of Active Learning with multiple annotators, the burden can be somewhat mitigated by actively acquiring labels of most informative data instances. However, multiple annotators with varying degrees of expertise poses new set of challenges in terms of quality of the label received and availability of the annotator. Due to limited amount of ground truth information addressing the variabilities of Activity of Daily Living (ADLs), activity recognition models using wearable and mobile devices are still not robust enough for real-world deployment. In this paper, we first propose an active learning combined deep model which updates its network parameters based on the optimization of a joint loss function. We then propose a novel annotator selection model by exploiting the relationships among the users while considering their heterogeneity with respect to their expertise, physical and spatial context. Our proposed model leverages model-free deep reinforcement learning in a partially observable environment setting to capture the action-reward interaction among multiple annotators. Our experiments in real-world settings exhibit that our active deep model converges to optimal accuracy with fewer labeled instances and achieves ~8% improvement in accuracy in fewer iterations.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133040318","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}
Data and data analysis are widely assumed to be the key part of the solution to healthcare systems' problems. Indeed, there are countless ways in which data can be converted into better medical diagnostic tools, more effective therapeutics, and improved productivity for clinicians. But while there is clearly great potential, some big challenges remain to make this all a reality, including making access to health data easier, addressing privacy and ethics concerns, and ensuring the clinical safety of "learning" systems. This talk illustrates what is possible in healthcare technology, and details key challenges that currently prevent this from becoming a reality.
{"title":"The Unreasonable Effectiveness, and Difficulty, of Data in Healthcare","authors":"Peter Lee","doi":"10.1145/3292500.3330645","DOIUrl":"https://doi.org/10.1145/3292500.3330645","url":null,"abstract":"Data and data analysis are widely assumed to be the key part of the solution to healthcare systems' problems. Indeed, there are countless ways in which data can be converted into better medical diagnostic tools, more effective therapeutics, and improved productivity for clinicians. But while there is clearly great potential, some big challenges remain to make this all a reality, including making access to health data easier, addressing privacy and ethics concerns, and ensuring the clinical safety of \"learning\" systems. This talk illustrates what is possible in healthcare technology, and details key challenges that currently prevent this from becoming a reality.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122966645","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}
Dimitris Spathis, S. S. Rodríguez, K. Farrahi, C. Mascolo, Jason Rentfrow
Smartphones have started to be used as self reporting tools for mental health state as they accompany individuals during their days and can therefore gather temporally fine grained data. However, the analysis of self reported mood data offers challenges related to non-homogeneity of mood assessment among individuals due to the complexity of the feeling and the reporting scales, as well as the noise and sparseness of the reports when collected in the wild. In this paper, we propose a new end-to-end ML model inspired by video frame prediction and machine translation, that forecasts future sequences of mood from previous self-reported moods collected in the real world using mobile devices. Contrary to traditional time series forecasting algorithms, our multi-task encoder-decoder recurrent neural network learns patterns from different users, allowing and improving the prediction for users with limited number of self-reports. Unlike traditional feature-based machine learning algorithms, the encoder-decoder architecture enables to forecast a sequence of future moods rather than one single step. Meanwhile, multi-task learning exploits some unique characteristics of the data (mood is bi-dimensional), achieving better results than when training single-task networks or other classifiers. Our experiments using a real-world dataset of 33,000 user-weeks revealed that (i) 3 weeks of sparsely reported mood is the optimal number to accurately forecast mood, (ii) multi-task learning models both dimensions of mood "valence and arousal" with higher accuracy than separate or traditional ML models, and (iii) mood variability, personality traits and day of the week play a key role in the performance of our model. We believe this work provides psychologists and developers of future mobile mental health applications with a ready-to-use and effective tool for early diagnosis of mental health issues at scale.
{"title":"Sequence Multi-task Learning to Forecast Mental Wellbeing from Sparse Self-reported Data","authors":"Dimitris Spathis, S. S. Rodríguez, K. Farrahi, C. Mascolo, Jason Rentfrow","doi":"10.1145/3292500.3330730","DOIUrl":"https://doi.org/10.1145/3292500.3330730","url":null,"abstract":"Smartphones have started to be used as self reporting tools for mental health state as they accompany individuals during their days and can therefore gather temporally fine grained data. However, the analysis of self reported mood data offers challenges related to non-homogeneity of mood assessment among individuals due to the complexity of the feeling and the reporting scales, as well as the noise and sparseness of the reports when collected in the wild. In this paper, we propose a new end-to-end ML model inspired by video frame prediction and machine translation, that forecasts future sequences of mood from previous self-reported moods collected in the real world using mobile devices. Contrary to traditional time series forecasting algorithms, our multi-task encoder-decoder recurrent neural network learns patterns from different users, allowing and improving the prediction for users with limited number of self-reports. Unlike traditional feature-based machine learning algorithms, the encoder-decoder architecture enables to forecast a sequence of future moods rather than one single step. Meanwhile, multi-task learning exploits some unique characteristics of the data (mood is bi-dimensional), achieving better results than when training single-task networks or other classifiers. Our experiments using a real-world dataset of 33,000 user-weeks revealed that (i) 3 weeks of sparsely reported mood is the optimal number to accurately forecast mood, (ii) multi-task learning models both dimensions of mood \"valence and arousal\" with higher accuracy than separate or traditional ML models, and (iii) mood variability, personality traits and day of the week play a key role in the performance of our model. We believe this work provides psychologists and developers of future mobile mental health applications with a ready-to-use and effective tool for early diagnosis of mental health issues at scale.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123982651","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}
Chengxi Zang, Peng Cui, Chaoming Song, Wenwu Zhu, Fei Wang
Pattern formation is a ubiquitous phenomenon that describes the generation of orderly outcomes by self-organization. In both physical society and online social media, patterns formed by social interactions are mainly driven by information flow. Despite an increasing number of studies aiming to understand the spreads of information flow, little is known about the geometry of these spreading patterns and how they were formed during the spreading. In this paper, by exploring 432 million information flow patterns extracted from a large-scale online social media dataset, we uncover a wide range of complex geometric patterns characterized by a three-dimensional metric space. In contrast, the existing understanding of spreading patterns are limited to fanning-out or narrow tree-like geometries. We discover three key ingredients that govern the formation of complex geometric patterns of information flow. As a result, we propose a stochastic process model incorporating these ingredients, demonstrating that it successfully reproduces the diverse geometries discovered from the empirical spreading patterns. Our discoveries provide a theoretical foundation for the microscopic mechanisms of information flow, potentially leading to wide implications for prediction, control and policy decisions in social media.
{"title":"Uncovering Pattern Formation of Information Flow","authors":"Chengxi Zang, Peng Cui, Chaoming Song, Wenwu Zhu, Fei Wang","doi":"10.1145/3292500.3330971","DOIUrl":"https://doi.org/10.1145/3292500.3330971","url":null,"abstract":"Pattern formation is a ubiquitous phenomenon that describes the generation of orderly outcomes by self-organization. In both physical society and online social media, patterns formed by social interactions are mainly driven by information flow. Despite an increasing number of studies aiming to understand the spreads of information flow, little is known about the geometry of these spreading patterns and how they were formed during the spreading. In this paper, by exploring 432 million information flow patterns extracted from a large-scale online social media dataset, we uncover a wide range of complex geometric patterns characterized by a three-dimensional metric space. In contrast, the existing understanding of spreading patterns are limited to fanning-out or narrow tree-like geometries. We discover three key ingredients that govern the formation of complex geometric patterns of information flow. As a result, we propose a stochastic process model incorporating these ingredients, demonstrating that it successfully reproduces the diverse geometries discovered from the empirical spreading patterns. Our discoveries provide a theoretical foundation for the microscopic mechanisms of information flow, potentially leading to wide implications for prediction, control and policy decisions in social media.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121241982","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}
The Multi-Armed Bandit (MAB) is a fundamental model capturing the dilemma between exploration and exploitation in sequential decision making. At every time step, the decision maker selects a set of arms and observes a reward from each of the chosen arms. In this paper, we present a variant of the problem, which we call the Scaling MAB (S-MAB): The goal of the decision maker is not only to maximize the cumulative rewards, i.e., choosing the arms with the highest expected reward, but also to decide how many arms to select so that, in expectation, the cost of selecting arms does not exceed the rewards. This problem is relevant to many real-world applications, e.g., online advertising, financial investments or data stream monitoring. We propose an extension of Thompson Sampling, which has strong theoretical guarantees and is reported to perform well in practice. Our extension dynamically controls the number of arms to draw. Furthermore, we combine the proposed method with ADWIN, a state-of-the-art change detector, to deal with non-static environments. We illustrate the benefits of our contribution via a real-world use case on predictive maintenance.
{"title":"Scaling Multi-Armed Bandit Algorithms","authors":"Edouard Fouché, Junpei Komiyama, Klemens Böhm","doi":"10.1145/3292500.3330862","DOIUrl":"https://doi.org/10.1145/3292500.3330862","url":null,"abstract":"The Multi-Armed Bandit (MAB) is a fundamental model capturing the dilemma between exploration and exploitation in sequential decision making. At every time step, the decision maker selects a set of arms and observes a reward from each of the chosen arms. In this paper, we present a variant of the problem, which we call the Scaling MAB (S-MAB): The goal of the decision maker is not only to maximize the cumulative rewards, i.e., choosing the arms with the highest expected reward, but also to decide how many arms to select so that, in expectation, the cost of selecting arms does not exceed the rewards. This problem is relevant to many real-world applications, e.g., online advertising, financial investments or data stream monitoring. We propose an extension of Thompson Sampling, which has strong theoretical guarantees and is reported to perform well in practice. Our extension dynamically controls the number of arms to draw. Furthermore, we combine the proposed method with ADWIN, a state-of-the-art change detector, to deal with non-static environments. We illustrate the benefits of our contribution via a real-world use case on predictive maintenance.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128545532","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}
Real-world entities' behaviors, associated with their side information, are often recorded over time as asynchronous event sequences. Such event sequences are the basis of many practical applications, neural spiking train study, earth quack prediction, crime analysis, infectious disease diffusion forecasting, condition-based preventative maintenance, information retrieval and behavior-based network analysis and services, etc. Temporal point process (TPP) is a principled mathematical tool for the modeling and learning of asynchronous event sequences, which captures the instantaneous happening rate of the events and the temporal dependency between historical and current events. TPP provides us with an interpretable model to describe the generative mechanism of event sequences, which is beneficial for event prediction and causality analysis. Recently, it has been shown that TPP has potentials to many machine learning and data science applications and can be combined with other cutting-edge machine learning techniques like deep learning, reinforcement learning, adversarial learning, and so on. We will start with an elementary introduction of TPP model, including the basic concepts of the model, the simulation method of event sequences; in the second part of the tutorial, we will introduce typical TPP models and their traditional learning methods; in the third part of the tutorial, we will discuss the recent progress on the modeling and learning of TPP, including neural network-based TPP models, generative adversarial networks (GANs) for TPP, and deep reinforcement learning of TPP. We will further talk about the practical application of TPP, including useful data augmentation methods for learning from imperfect observations, typical applications and examples like healthcare and industry maintenance, and existing open source toolboxes.
{"title":"Modeling and Applications for Temporal Point Processes","authors":"Junchi Yan, Hongteng Xu, Liangda Li","doi":"10.1145/3292500.3332298","DOIUrl":"https://doi.org/10.1145/3292500.3332298","url":null,"abstract":"Real-world entities' behaviors, associated with their side information, are often recorded over time as asynchronous event sequences. Such event sequences are the basis of many practical applications, neural spiking train study, earth quack prediction, crime analysis, infectious disease diffusion forecasting, condition-based preventative maintenance, information retrieval and behavior-based network analysis and services, etc. Temporal point process (TPP) is a principled mathematical tool for the modeling and learning of asynchronous event sequences, which captures the instantaneous happening rate of the events and the temporal dependency between historical and current events. TPP provides us with an interpretable model to describe the generative mechanism of event sequences, which is beneficial for event prediction and causality analysis. Recently, it has been shown that TPP has potentials to many machine learning and data science applications and can be combined with other cutting-edge machine learning techniques like deep learning, reinforcement learning, adversarial learning, and so on. We will start with an elementary introduction of TPP model, including the basic concepts of the model, the simulation method of event sequences; in the second part of the tutorial, we will introduce typical TPP models and their traditional learning methods; in the third part of the tutorial, we will discuss the recent progress on the modeling and learning of TPP, including neural network-based TPP models, generative adversarial networks (GANs) for TPP, and deep reinforcement learning of TPP. We will further talk about the practical application of TPP, including useful data augmentation methods for learning from imperfect observations, typical applications and examples like healthcare and industry maintenance, and existing open source toolboxes.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115359393","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}
Lisi Chen, Shuo Shang, Christian S. Jensen, Bin Yao, Zhiwei Zhang, Ling Shao
With the increasing availability of moving-object tracking data, use of this data for route search and recommendation is increasingly important. To this end, we propose a novel parallel split-and-combine approach to enable route search by locations (RSL-Psc). Given a set of routes, a set of places to visit O, and a threshold θ, we retrieve the route composed of sub-routes that (i) has similarity to O no less than θ and (ii) contains the minimum number of sub-route combinations. The resulting functionality targets a broad range of applications, including route planning and recommendation, ridesharing, and location-based services in general. To enable efficient and effective RSL-Psc computation on massive route data, we develop novel search space pruning techniques and enable use of the parallel processing capabilities of modern processors. Specifically, we develop two parallel algorithms, Fully-Split Parallel Search (FSPS) and Group-Split Parallel Search (GSPS). We divide the route split-and-combine task into ∑k=0 M S(|O|,k+1) sub-tasks, where M is the maximum number of combinations and S(⋅) is the Stirling number of the second kind. In each sub-task, we use network expansion and exploit spatial similarity bounds for pruning. The algorithms split candidate routes into sub-routes and combine them to construct new routes. The sub-tasks are independent and are performed in parallel. Extensive experiments with real data offer insight into the performance of the algorithms, indicating that our RSL-Psc problem can generate high-quality results and that the two algorithms are capable of achieving high efficiency and scalability.
随着移动目标跟踪数据的可用性越来越高，使用这些数据进行路线搜索和推荐变得越来越重要。为此，我们提出了一种新的并行拆分合并方法来实现按位置的路由搜索(RSL-Psc)。给定一组路由，一组访问O的地点和一个阈值θ，我们检索由子路由组成的路由，其中(i)与O的相似性不小于θ， (ii)包含最小数量的子路由组合。由此产生的功能针对广泛的应用，包括路线规划和推荐，乘车共享和基于位置的服务。为了在大量路由数据上实现高效的RSL-Psc计算，我们开发了新的搜索空间修剪技术，并启用了现代处理器的并行处理能力。具体来说，我们开发了两种并行算法，全分割并行搜索(FSPS)和组分割并行搜索(GSPS)。我们将路径分割合并任务划分为∑k=0 M S(|O|，k+1)个子任务，其中M为最大组合数，S(⋅)为第二类斯特林数。在每个子任务中，我们使用网络扩展和利用空间相似界进行修剪。该算法将候选路由分解成子路由，并将它们组合成新的路由。子任务是独立的，并且并行执行。大量的真实数据实验提供了对算法性能的深入了解，表明我们的RSL-Psc问题可以产生高质量的结果，并且两种算法能够实现高效率和可扩展性。
{"title":"Effective and Efficient Reuse of Past Travel Behavior for Route Recommendation","authors":"Lisi Chen, Shuo Shang, Christian S. Jensen, Bin Yao, Zhiwei Zhang, Ling Shao","doi":"10.1145/3292500.3330835","DOIUrl":"https://doi.org/10.1145/3292500.3330835","url":null,"abstract":"With the increasing availability of moving-object tracking data, use of this data for route search and recommendation is increasingly important. To this end, we propose a novel parallel split-and-combine approach to enable route search by locations (RSL-Psc). Given a set of routes, a set of places to visit O, and a threshold θ, we retrieve the route composed of sub-routes that (i) has similarity to O no less than θ and (ii) contains the minimum number of sub-route combinations. The resulting functionality targets a broad range of applications, including route planning and recommendation, ridesharing, and location-based services in general. To enable efficient and effective RSL-Psc computation on massive route data, we develop novel search space pruning techniques and enable use of the parallel processing capabilities of modern processors. Specifically, we develop two parallel algorithms, Fully-Split Parallel Search (FSPS) and Group-Split Parallel Search (GSPS). We divide the route split-and-combine task into ∑k=0 M S(|O|,k+1) sub-tasks, where M is the maximum number of combinations and S(⋅) is the Stirling number of the second kind. In each sub-task, we use network expansion and exploit spatial similarity bounds for pruning. The algorithms split candidate routes into sub-routes and combine them to construct new routes. The sub-tasks are independent and are performed in parallel. Extensive experiments with real data offer insight into the performance of the algorithms, indicating that our RSL-Psc problem can generate high-quality results and that the two algorithms are capable of achieving high efficiency and scalability.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114418382","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}
Ho-Yong Lee, Jinbae Im, Seongwon Jang, H. Cho, Sehee Chung
This paper proposes a recommender system to alleviate the cold-start problem that can estimate user preferences based on only a small number of items. To identify a user's preference in the cold state, existing recommender systems, such as Netflix, initially provide items to a user; we call those items evidence candidates. Recommendations are then made based on the items selected by the user. Previous recommendation studies have two limitations: (1) the users who consumed a few items have poor recommendations and (2) inadequate evidence candidates are used to identify user preferences. We propose a meta-learning-based recommender system called MeLU to overcome these two limitations. From meta-learning, which can rapidly adopt new task with a few examples, MeLU can estimate new user's preferences with a few consumed items. In addition, we provide an evidence candidate selection strategy that determines distinguishing items for customized preference estimation. We validate MeLU with two benchmark datasets, and the proposed model reduces at least 5.92% mean absolute error than two comparative models on the datasets. We also conduct a user study experiment to verify the evidence selection strategy.
{"title":"MeLU","authors":"Ho-Yong Lee, Jinbae Im, Seongwon Jang, H. Cho, Sehee Chung","doi":"10.1145/3292500.3330859","DOIUrl":"https://doi.org/10.1145/3292500.3330859","url":null,"abstract":"This paper proposes a recommender system to alleviate the cold-start problem that can estimate user preferences based on only a small number of items. To identify a user's preference in the cold state, existing recommender systems, such as Netflix, initially provide items to a user; we call those items evidence candidates. Recommendations are then made based on the items selected by the user. Previous recommendation studies have two limitations: (1) the users who consumed a few items have poor recommendations and (2) inadequate evidence candidates are used to identify user preferences. We propose a meta-learning-based recommender system called MeLU to overcome these two limitations. From meta-learning, which can rapidly adopt new task with a few examples, MeLU can estimate new user's preferences with a few consumed items. In addition, we provide an evidence candidate selection strategy that determines distinguishing items for customized preference estimation. We validate MeLU with two benchmark datasets, and the proposed model reduces at least 5.92% mean absolute error than two comparative models on the datasets. We also conduct a user study experiment to verify the evidence selection strategy.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114492515","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}
Lyft's mission is to improve people's lives with the world's best transportation. Self driving vehicles have the potential to deliver unprecedented improvements to safety and quality, at a price and convenience that challenges traditional models of vehicle ownership. A combination of hardware, software, and knowledge technologies are needed to build self-driving cars. In this talk, I'll present the core problems in self-driving and how recent advances in computer vision, robotics, and machine learning are powering this revolution. The car is carefully designed with a variety of sensors that complement each other to address a wide variety of driving scenarios. Sensor fusion bring all of these signals together into an interpretable AI engine comprising of perception, prediction, planning, and controls. For example, deep learning models and large scale machine learning have closed the gap between human and machine perception. In contrast, predicting the behavior of other humans and effectively planning and negotiating maneuvers continue to be hard problems. Combining AI technologies with deep knowledge about the real world is key to addressing these.
{"title":"Building a Better Self-Driving Car: Hardware, Software, and Knowledge","authors":"K. Chellapilla","doi":"10.1145/3292500.3340409","DOIUrl":"https://doi.org/10.1145/3292500.3340409","url":null,"abstract":"Lyft's mission is to improve people's lives with the world's best transportation. Self driving vehicles have the potential to deliver unprecedented improvements to safety and quality, at a price and convenience that challenges traditional models of vehicle ownership. A combination of hardware, software, and knowledge technologies are needed to build self-driving cars. In this talk, I'll present the core problems in self-driving and how recent advances in computer vision, robotics, and machine learning are powering this revolution. The car is carefully designed with a variety of sensors that complement each other to address a wide variety of driving scenarios. Sensor fusion bring all of these signals together into an interpretable AI engine comprising of perception, prediction, planning, and controls. For example, deep learning models and large scale machine learning have closed the gap between human and machine perception. In contrast, predicting the behavior of other humans and effectively planning and negotiating maneuvers continue to be hard problems. Combining AI technologies with deep knowledge about the real world is key to addressing these.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115242712","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}
There have long been connections between statistical mechanics and neural networks, but in recent decades these connections have withered. However, in light of recent failings of statistical learning theory and stochastic optimization theory to describe, even qualitatively, many properties of production-quality neural network models, researchers have revisited ideas from the statistical mechanics of neural networks. This tutorial will provide an overview of the area; it will go into detail on how connections with random matrix theory and heavy-tailed random matrix theory can lead to a practical phenomenological theory for large-scale deep neural networks; and it will describe future directions.
{"title":"Statistical Mechanics Methods for Discovering Knowledge from Modern Production Quality Neural Networks","authors":"Charles H. Martin, Michael W. Mahoney","doi":"10.1145/3292500.3332294","DOIUrl":"https://doi.org/10.1145/3292500.3332294","url":null,"abstract":"There have long been connections between statistical mechanics and neural networks, but in recent decades these connections have withered. However, in light of recent failings of statistical learning theory and stochastic optimization theory to describe, even qualitatively, many properties of production-quality neural network models, researchers have revisited ideas from the statistical mechanics of neural networks. This tutorial will provide an overview of the area; it will go into detail on how connections with random matrix theory and heavy-tailed random matrix theory can lead to a practical phenomenological theory for large-scale deep neural networks; and it will describe future directions.","PeriodicalId":186134,"journal":{"name":"Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114790757","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：