scholarly journals Popularity, novelty and relevance in point of interest recommendation: an experimental analysis

2021 ◽  
Author(s):  
David Massimo ◽  
Francesco Ricci
Keyword(s):  
Reinforcement Learning ◽  
Experimental Analysis ◽  
State Of The Art ◽  
The Other ◽  
Nearest Neighbour ◽  
Inverse Reinforcement Learning ◽  
Point Of Interest ◽  
On Line ◽  
Observed Behaviour ◽  
The Cost

AbstractRecommender Systems (RSs) are often assessed in off-line settings by measuring the system precision in predicting the observed user’s ratings or choices. But, when a precise RS is on-line, the generated recommendations can be perceived as marginally useful because lacking novelty. The underlying problem is that it is hard to build an RS that can correctly generalise, from the analysis of user’s observed behaviour, and can identify the essential characteristics of novel and yet relevant recommendations. In this paper we address the above mentioned issue by considering four RSs that try to excel on different target criteria: precision, relevance and novelty. Two state of the art RSs called and follow a classical Nearest Neighbour approach, while the other two, and are based on Inverse Reinforcement Learning. and optimise precision, tries to identify the characteristics of POIs that make them relevant, and , a novel RS here introduced, is similar to but it also tries to recommend popular POIs. In an off-line experiment we discover that the recommendations produced by and optimise precision essentially by recommending quite popular POIs. can be tuned to achieve a desired level of precision at the cost of losing part of the best capability of to generate novel and yet relevant recommendations. In the on-line study we discover that the recommendations of and are liked more than those produced by . The rationale of that was found in the large percentage of novel recommendations produced by , which are difficult to appreciate. However, excels in recommending items that are both novel and liked by the users.

N and P on-line meters: requirements, maintenance and stability

1996 ◽  
Vol 33 (1) ◽  
pp. 147-157 ◽  
Author(s):  
Henrik A. Thomsen ◽  
Kenneth Kisbye
Keyword(s):  
Sample Preparation ◽  
Response Time ◽  
Ammonium Nitrate ◽  
State Of The Art ◽  
List Price ◽  
User Friendliness ◽  
On Line ◽  
The Cost ◽  
Preparation Design

State-of-the-art on-line meters for determination of ammonium, nitrate and phosphate are presented. The on-line meters employ different measuring principles and are available in many different designs differing with respect to size, calibration and cleaning principle, user-friendliness, response time, reagent and sample consumption. A study of Danish experiences on several plants has been conducted. The list price of an on-line meter is between USD 8000 and USD 35,000. To this should be added the cost of sample preparation, design, installation and running-in. The yearly operating for one meter are in the range of USD 200-2500 and the manpower consumption is in the range of 1-5 hours/month. The accuracy obtained is only slightly smaller than the accuracy on collaborative laboratory analyses, which is sufficient for most control purposes.

scholarly journals Dialogue Generation: From Imitation Learning to Inverse Reinforcement Learning

2019 ◽  
Vol 33 ◽  
pp. 6722-6729 ◽  
Author(s):  
Ziming Li ◽  
Julia Kiseleva ◽  
Maarten De Rijke
Keyword(s):  
Reinforcement Learning ◽  
State Of The Art ◽  
The State ◽  
Experimental Results ◽  
Imitation Learning ◽  
Local Optimum ◽  
Inverse Reinforcement Learning ◽  
High Quality ◽  
Overall Performance

The performance of adversarial dialogue generation models relies on the quality of the reward signal produced by the discriminator. The reward signal from a poor discriminator can be very sparse and unstable, which may lead the generator to fall into a local optimum or to produce nonsense replies. To alleviate the first problem, we first extend a recently proposed adversarial dialogue generation method to an adversarial imitation learning solution. Then, in the framework of adversarial inverse reinforcement learning, we propose a new reward model for dialogue generation that can provide a more accurate and precise reward signal for generator training. We evaluate the performance of the resulting model with automatic metrics and human evaluations in two annotation settings. Our experimental results demonstrate that our model can generate more high-quality responses and achieve higher overall performance than the state-of-the-art.

scholarly journals Object Affordance Driven Inverse Reinforcement Learning Through Conceptual Abstraction and Advice

2018 ◽  
Vol 9 (1) ◽  
pp. 277-294 ◽  
Author(s):  
Rupam Bhattacharyya ◽  
Shyamanta M. Hazarika
Keyword(s):  
Reinforcement Learning ◽  
High Dimensional ◽  
Inverse Reinforcement Learning ◽  
Intent Recognition ◽  
Reward Function ◽  
Object Affordances ◽  
Learning Agent ◽  
Markov Decision ◽  
Observed Behaviour ◽  
Object Affordance

Abstract Within human Intent Recognition (IR), a popular approach to learning from demonstration is Inverse Reinforcement Learning (IRL). IRL extracts an unknown reward function from samples of observed behaviour. Traditional IRL systems require large datasets to recover the underlying reward function. Object affordances have been used for IR. Existing literature on recognizing intents through object affordances fall short of utilizing its true potential. In this paper, we seek to develop an IRL system which drives human intent recognition along with the capability to handle high dimensional demonstrations exploiting the capability of object affordances. An architecture for recognizing human intent is presented which consists of an extended Maximum Likelihood Inverse Reinforcement Learning agent. Inclusion of Symbolic Conceptual Abstraction Engine (SCAE) along with an advisor allows the agent to work on Conceptually Abstracted Markov Decision Process. The agent recovers object affordance based reward function from high dimensional demonstrations. This function drives a Human Intent Recognizer through identification of probable intents. Performance of the resulting system on the standard CAD-120 dataset shows encouraging result.

Large-scale cost function learning for path planning using deep inverse reinforcement learning

2017 ◽  
Vol 36 (10) ◽  
pp. 1073-1087 ◽  
Author(s):  
Markus Wulfmeier ◽  
Dushyant Rao ◽  
Dominic Zeng Wang ◽  
Peter Ondruska ◽  
Ingmar Posner
Keyword(s):  
Reinforcement Learning ◽  
Large Scale ◽  
Cost Model ◽  
High Capacity ◽  
Urban Environments ◽  
Training Dataset ◽  
Inverse Reinforcement Learning ◽  
Dataset Size ◽  
And Performance ◽  
The Cost

We present an approach for learning spatial traversability maps for driving in complex, urban environments based on an extensive dataset demonstrating the driving behaviour of human experts. The direct end-to-end mapping from raw input data to cost bypasses the effort of manually designing parts of the pipeline, exploits a large number of data samples, and can be framed additionally to refine handcrafted cost maps produced based on manual hand-engineered features. To achieve this, we introduce a maximum-entropy-based, non-linear inverse reinforcement learning (IRL) framework which exploits the capacity of fully convolutional neural networks (FCNs) to represent the cost model underlying driving behaviours. The application of a high-capacity, deep, parametric approach successfully scales to more complex environments and driving behaviours, while at deployment being run-time independent of training dataset size. After benchmarking against state-of-the-art IRL approaches, we focus on demonstrating scalability and performance on an ambitious dataset collected over the course of 1 year including more than 25,000 demonstration trajectories extracted from over 120 km of urban driving. We evaluate the resulting cost representations by showing the advantages over a carefully, manually designed cost map and furthermore demonstrate its robustness towards systematic errors by learning accurate representations even in the presence of calibration perturbations. Importantly, we demonstrate that a manually designed cost map can be refined to more accurately handle corner cases that are scarcely seen in the environment, such as stairs, slopes and underpasses, by further incorporating human priors into the training framework.

scholarly journals Bayes-Adaptive Monte-Carlo Planning and Learning for Goal-Oriented Dialogues

2020 ◽  
Vol 34 (05) ◽  
pp. 7994-8001
Author(s):  
Youngsoo Jang ◽  
Jongmin Lee ◽  
Kee-Eung Kim
Keyword(s):  
Reinforcement Learning ◽  
Task Performance ◽  
State Of The Art ◽  
Search Space ◽  
Difficult Problem ◽  
The Other ◽  
Adaptive Planning ◽  
Adaptive Monte Carlo ◽  
Planning Algorithm ◽  
Dialogue Planning

We consider a strategic dialogue task, where the ability to infer the other agent's goal is critical to the success of the conversational agent. While this problem can be naturally formulated as Bayesian planning, it is known to be a very difficult problem due to its enormous search space consisting of all possible utterances. In this paper, we introduce an efficient Bayes-adaptive planning algorithm for goal-oriented dialogues, which combines RNN-based dialogue generation and MCTS-based Bayesian planning in a novel way, leading to robust decision-making under the uncertainty of the other agent's goal. We then introduce reinforcement learning for the dialogue agent that uses MCTS as a strong policy improvement operator, casting reinforcement learning as iterative alternation of planning and supervised-learning of self-generated dialogues. In the experiments, we demonstrate that our Bayes-adaptive dialogue planning agent significantly outperforms the state-of-the-art in a negotiation dialogue domain. We also show that reinforcement learning via MCTS further improves end-task performance without diverging from human language.

scholarly journals Constrained Bayesian Reinforcement Learning via Approximate Linear Programming

2017 ◽  
Author(s):  
Jongmin Lee ◽  
Youngsoo Jang ◽  
Pascal Poupart ◽  
Kee-Eung Kim
Keyword(s):  
Linear Programming ◽  
Reinforcement Learning ◽  
State Of The Art ◽  
Learning Agent ◽  
Risk Sensitive ◽  
State Controller ◽  
Finite State ◽  
Bayesian Reinforcement Learning ◽  
Approximate Linear Programming ◽  
The Cost

In this paper, we consider the safe learning scenario where we need to restrict the exploratory behavior of a reinforcement learning agent. Specifically, we treat the problem as a form of Bayesian reinforcement learning in an environment that is modeled as a constrained MDP (CMDP) where the cost function penalizes undesirable situations. We propose a model-based Bayesian reinforcement learning (BRL) algorithm for such an environment, eliciting risk-sensitive exploration in a principled way. Our algorithm efficiently solves the constrained BRL problem by approximate linear programming, and generates a finite state controller in an off-line manner. We provide theoretical guarantees and demonstrate empirically that our approach outperforms the state of the art.

scholarly journals Toward Self-Driving Bicycles Using State-of-the-Art Deep Reinforcement Learning Algorithms

Symmetry ◽  
2019 ◽  
Vol 11 (2) ◽  
pp. 290 ◽  
Author(s):  
SeungYoon Choi ◽  
Tuyen Le ◽  
Quang Nguyen ◽  
Md Layek ◽  
SeungGwan Lee ◽  
...  
Keyword(s):  
Reinforcement Learning ◽  
Deep Neural Network ◽  
Learning Algorithm ◽  
State Of The Art ◽  
The Other ◽  
Gradient Algorithm ◽  
Reward Function ◽  
Policy Gradient ◽  
Policy Optimization ◽  
Start Location

In this paper, we propose a controller for a bicycle using the DDPG (Deep Deterministic Policy Gradient) algorithm, which is a state-of-the-art deep reinforcement learning algorithm. We use a reward function and a deep neural network to build the controller. By using the proposed controller, a bicycle can not only be stably balanced but also travel to any specified location. We confirm that the controller with DDPG shows better performance than the other baselines such as Normalized Advantage Function (NAF) and Proximal Policy Optimization (PPO). For the performance evaluation, we implemented the proposed algorithm in various settings such as fixed and random speed, start location, and destination location.

scholarly journals Um agrupamento de modelos conexonistas por meio de sinapses artificiais e suas aplicações no mercado de criptomoedas

2020 ◽  
Author(s):  
◽  
C. O. Vilão Júnior
Keyword(s):  
Neural Networks ◽  
Reinforcement Learning ◽  
State Of The Art ◽  
A Priori ◽  
The Other ◽  
Biological Inspiration ◽  
Deep Networks ◽  
Ensemble Algorithms ◽  
The Brain ◽  
Better Than

This work proposes an algorithm, named CMEAS, has biological inspiration focused on the way that the growth of neuronal axons reaches their synaptic destination in other neural networks. This growth follows specific pathways in the brain of animals defined by certain proteins. CMEAS was developed to group two convolutional neural networks, trained a priori on two topics that simultaneously influence the cryptocurrency market, such as news and prices. The means by which networks are grouped occurs using connections external to the original networks to connect to the internal neurons of each network. Two strands were proposed in order to train CMEAS, being one with supervised learning and the other with reinforcement learning. The results confirmed by the Wilcoxon tests demonstrate that the CMEAS had a better profit factor and a higher sharpe index in the experiments in relation to the classic ensemble algorithms through voting and stand-alone deep networks, the algorithm was also superior in all the metrics of the buy and hold strategy, in addition, the algorithm obtained similar results, however, better than those of CNN-LSTM considered state of the art, given the metrics used

Reading About Persons: The Effects of Conjunctions on the Mental Representation of Person Descriptions

2008 ◽  
Vol 67 (1) ◽  
pp. 5-18 ◽  
Author(s):  
Sabine Krolak-Schwerdt ◽  
Margret Wintermantel ◽  
Nadine Junker ◽  
Julia Kneer
Keyword(s):  
Mental Representation ◽  
The Other ◽  
Person Memory ◽  
Causal Connections ◽  
Single Person ◽  
On Line

Three experiments investigated the processing of person descriptions that consisted of a number of statements about the characteristics of a person. In one condition, each statement referred to a single person attribute and in the other condition, causal and additive conjunctions to verbally link the statements were introduced. Evidence was found that the introduction of verbal links enhanced participants’ memory about the characteristics of the described person. On-line measures of processing showed that the comprehension of person information was strongly facilitated by the introduction of verbal links. Furthermore, the results were due to the introduction of causal connections between person attributes. These findings are discussed in terms of their implications for models of person memory and representation.

Ensemble Inverse Reinforcement Learning from Semi-Expert Agents

2017 ◽  
Vol 137 (4) ◽  
pp. 667-673
Author(s):  
Shinji Tomita ◽  
Fumiya Hamatsu ◽  
Tomoki Hamagami
Keyword(s):  
Reinforcement Learning ◽  
Inverse Reinforcement Learning
Sign in / Sign up

Export Citation Format

Share Document