Language Assistance in Reinforcement Learning in Dynamic Environments

Language is an intuitive and effective way for humans to communicate. Large Language Models (LLMs) can interpret and respond well to language. However, their use in deep reinforcement learning is limited as they are sample inefficient. State-of-the-art deep reinforcement learning algorithms are more sample efficient but cannot understand...

Activity Progress Prediction: Is there progress in video progress prediction methods?

In this paper, we investigate the behaviour of current progress prediction methods on the currently used benchmark datasets. We show that the progress prediction methods can fail to extract useful information from visual data on these datasets. Moreover, when the methods fail to extract visual information, memory-based methods adopt a frame...

Prioritizing states with action sensitive return in experience replay

Experience replay for off-policy reinforcement learning has been shown to improve sample efficiency and stabilize training. However, typical uniformly sampled replay includes many irrelevant samples for the agent to reach good performance. We introduce Action Sensitive Experience Replay (ASER), a method to prioritize samples in the replay buffer...

Multi-objective Deep Reinforcement Learning for predictive maintenance of road networks

Operation and maintenance of the built environment have a major effect on socioeconomic stability and sustainability. A significant part of our built world approaches or has well exceeded its designated structural life. As engineers, we need to find efficient ways to extend this life while maintaining acceptable levels of safety and performance....

Vanishing empirical variance in randomly initialized networks

Neural networks are commonly initialized to keep the theoretical variance of the hidden pre-activations constant, in order to avoid the vanishing and exploding gradient problem. Though this condition is necessary to train very deep networks, numerous analyses showed that it is not sufficient. We explain this fact by analyzing the behavior of the...

Learning to search & track dynamic targets with graph representations

Autonomous robots have been widely applied to search and rescue missions for information gathering about target locations. This process needs to be continuously replanned based on new observations in the environment. For dynamic targets, the robot needs to not only discover them but also keep tracking their positions. Previous works focus on...

The effect of EHOP on the writing of Program Analyzers

Effect Handler Oriented Programming is a promising new programming paradigm, delivering separation of of concerns with regards to side effects in an otherwise functional environment.<br/>This paper discusses the applicability of this new paradigm to static code analysis programs. <br/>Different code analyzers often have many similar, if not...

Effect Handler Oriented Programming for Data Processing Applications

Effect handler oriented programming or EHOP for short, is a new programming paradigm aiming to achieve separation of concerns in code which will lead to modular, readable and maintainable code. Since EHOP is significantly new, it is important to assess and compare it against traditional, commonly used paradigms in order to see if a wider...

Applicability of Effect Handler Oriented Programming (EHOP) for Text-Based Game Development

Effect handler oriented programming (EHOP) is a recently proposed programming paradigm, which aims to provide a separation of concerns by isolating the handling of side-effects from the main application logic. Nowadays, as the core concepts behind EHOP are being added to more and more programming languages, it is evident that EHOP is slowly but...

Programming With Effects: Improving the quality of code in IO intensive applications through effect handlers

Effect handler orient programming (EHOP) is a recently proposed programming paradigm that aims to provide a high-level abstraction in code. Using this paradigm, programmers are able to define operations as an effect, which are implemented by an effect handler. Functions can then use effects, allowing the effect operations to be used in this...

The effect of anthropomorphism in conversational agents

Conversational agents are being used more often. Many conversational agents have been designed to make them more humanlike to increase the engagement of the user. This research is looking into a couple of aspects of anthropomorphism to determine whether these aspects improve the engagement of the users. These aspects are typos with time delay...

Studying the Anthropomorphic Visual Cues in Conversational Agents

With the increase of anthropomorphic conversational agents, the research on the effects of anthropomorphism in conversational agents is on the rise. However the studies cannot seem to agree on many of the raised questions and the research on the proper design of conversational agents is still lacking. This research paper attempted to contribute...

An Exploratory Study on Conversational Agents Using Dynamic Conversation Styles

This study aims to examine the effect of dynamically aligning the conversation style to the user’s preference in a conversational agent environment. We suggest a technique to identify the user’s preferred conversation style, by scoring statements based on their writing style. We designed a within-subject experiment intended to measure the user...

Automatically Testing a Conversational Crowd Computing Platform

The rise in the use of crowd computing platforms led to the birth of Dandelion, a conversational crowd computing platform developed at TU Delft with the main goals being to connect students with researchers and to allow students to report on their well-being by using a friendly interface. Dandelion was tested manually up to the time of drafting...

Conversation Workflows with Micro-breaks in Conversational Crowdsourcing

With the growing popularity of conversational interfaces, new use-cases in which conversation may be an effective medium for human-computer interaction are explored. Recent works have shown how conversational interfaces can improve worker engagement in micro-task crowdsourcing and that task performance is comparable to web-based interfaces. This...