Reinforcement Learning with External Knowledge and Two-Stage Q-functions for Predicting Popular Reddit Threads
نویسندگان
چکیده
This paper addresses the problem of predicting popularity of comments in an online discussion forum using reinforcement learning, particularly addressing two challenges that arise from having natural language state and action spaces. First, the state representation, which characterizes the history of comments tracked in a discussion at a particular point, is augmented to incorporate the global context represented by discussions on world events available in an external knowledge source. Second, a two-stage Q-learning framework is introduced, making it feasible to search the combinatorial action space while also accounting for redundancy among sub-actions. We experiment with five Reddit communities, showing that the two methods improve over previous reported results on this task.
منابع مشابه
Deep Reinforcement Learning with a Combinatorial Action Space for Predicting Popular Reddit Threads
We introduce an online popularity prediction and tracking task as a benchmark task for reinforcement learning with a combinatorial, natural language action space. A specified number of discussion threads predicted to be popular are recommended, chosen from a fixed window of recent comments to track. Novel deep reinforcement learning architectures are studied for effective modeling of the value ...
متن کاملAn Online Q-learning Based Multi-Agent LFC for a Multi-Area Multi-Source Power System Including Distributed Energy Resources
This paper presents an online two-stage Q-learning based multi-agent (MA) controller for load frequency control (LFC) in an interconnected multi-area multi-source power system integrated with distributed energy resources (DERs). The proposed control strategy consists of two stages. The first stage is employed a PID controller which its parameters are designed using sine cosine optimization (SCO...
متن کاملDeep Reinforcement Learning with a Combinatorial Action Space for Predicting and Tracking Popular Discussion Threads
We introduce an online popularity prediction and tracking task as a benchmark task for reinforcement learning with a combinatorial, natural language action space. A specified number of discussion threads predicted to be popular are recommended, chosen from a fixed window of recent comments to track. Novel deep reinforcement learning architectures are studied for effective modeling of the value ...
متن کاملReinforcement Q-Learning and ILC with Self-Tuning Learning Rate for Contour Following Accuracy Improvement of Biaxial Motion Stage
Abstract—Biaxial motion stages are commonly used in precision motion applications. However, the contour following accuracy of a biaxial motion stage often suffers from system nonlinearities and external disturbances. To deal with the above-mentioned problem, a control scheme consisting of a reinforcement Q-learning controller with a self-tuning learning rate and two iterative learning controlle...
متن کاملHierarchical Functional Concepts for Knowledge Transfer among Reinforcement Learning Agents
This article introduces the notions of functional space and concept as a way of knowledge representation and abstraction for Reinforcement Learning agents. These definitions are used as a tool of knowledge transfer among agents. The agents are assumed to be heterogeneous; they have different state spaces but share a same dynamic, reward and action space. In other words, the agents are assumed t...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- CoRR
دوره abs/1704.06217 شماره
صفحات -
تاریخ انتشار 2017