Tobias Huber
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Tobias Huber has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Elisabeth André
Universitat Augsburg, Augsburg, Germany
12 shared publications - Cristina Conati
University of British Columbia
4 shared publications Elisabeth André
Chair for Human-Centered Artificial Intelligence, University of Augsburg, Augsburg, Germany
4 shared publications- Ruben Schlagowski
Chair for Human-Centered Artificial Intelligence, University of Augsburg, DE
4 shared publications - Katharina Weitz
Fraunhofer Institute for Telecommunications, Heinrich Hertz Institute, Berlin, Germany
3 shared publications - Silvan Mertes
Technical University of Applied Sciences Augsburg, Augsburg, Germany
3 shared publications
Latest Publications
- Open Access
- 1 August 2024
- Published by International Joint Conferences on Artificial Intelligence
- Vol. 1, 467-475
Relevant Irrelevance: Generating Alterfactual Explanations for Image Classifiers
In this paper, we demonstrate the feasibility of alterfactual explanations for black box image classifiers. Traditional explanation mechanisms from the field of Counterfactual Thinking are a widely-used paradigm for Explainable Artificial Intelligence (XAI), as they follow a natural way of reasoning that humans are familiar with. However, most common approaches from this field are based on communicating information about features or characteristics that are especially important for an AI's decision. However, to fully understand a decision, not only knowledge about relevant features is needed, but the awareness of irrelevant information also highly contributes to the creation of a user's mental model of an AI system. To this end, a novel approach for explaining AI systems called alterfactual explanations was recently proposed on a conceptual level. It is based on showing an alternative reality where irrelevant features of an AI's input are altered. By doing so, the user directly sees which input data characteristics can change arbitrarily without influencing the AI's decision. In this paper, we show for the first time that it is possible to apply this idea to black box models based on neural networks. To this end, we present a GAN-based approach to generate these alterfactual explanations for binary image classifiers. Further, we present a user study that gives interesting insights on how alterfactual explanations can complement counterfactual explanations.
- 1 August 2024
- Published by International Joint Conferences on Artificial Intelligence
- Vol. 1, 467-475
Open AccessRelevant Irrelevance: Generating Alterfactual Explanations for Image Classifiers
In this paper, we demonstrate the feasibility of alterfactual explanations for black box image classifiers. Traditional explanation mechanisms from the field of Counterfactual Thinking are a widely-used paradigm for Explainable Artificial Intelligence (XAI), as they follow a natural way of reasoning that humans are familiar with. However, most common approaches from this field are based on communicating information about features or characteristics that are especially important for an AI's decision. However, to fully understand a decision, not only knowledge about relevant features is needed, but the awareness of irrelevant information also highly contributes to the creation of a user's mental model of an AI system. To this end, a novel approach for explaining AI systems called alterfactual explanations was recently proposed on a conceptual level. It is based on showing an alternative reality where irrelevant features of an AI's input are altered. By doing so, the user directly sees which input data characteristics can change arbitrarily without influencing the AI's decision. In this paper, we show for the first time that it is possible to apply this idea to black box models based on neural networks. To this end, we present a GAN-based approach to generate these alterfactual explanations for binary image classifiers. Further, we present a user study that gives interesting insights on how alterfactual explanations can complement counterfactual explanations.
- Open Access
- 21 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-12
From a Social POV: The Impact of Point of View on Player Behavior, Engagement, and Experience in a Serious Social Simulation Game
Multiplayer games with social aspects vary widely regarding client design, e.g., point of view or camera perspective. While design paradigms usually arise from gold standards that are set by previously successful games in the industry, the impact of such paradigms is under-researched for games that serve as scientific instruments, e.g., to research social behavior. Intending to investigate how such games should be designed, we built two multiplayer clients with the same game logic, one using a first-person point of view, while the other includes a top-down camera perspective. Then, we conducted an online user study in which players tested these game clients in extensive multiplayer sessions. Analyzing speech time, in-game logs, questionnaires, and qualitative feedback, we look at the perspectives’ impact on player behavior, engagement, and game experience in a scientific or "serious games" context. In addition, we have made our designed game UNISON and both clients available as open source to facilitate future empirical social science research.
- 21 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-12
Open AccessFrom a Social POV: The Impact of Point of View on Player Behavior, Engagement, and Experience in a Serious Social Simulation Game
Multiplayer games with social aspects vary widely regarding client design, e.g., point of view or camera perspective. While design paradigms usually arise from gold standards that are set by previously successful games in the industry, the impact of such paradigms is under-researched for games that serve as scientific instruments, e.g., to research social behavior. Intending to investigate how such games should be designed, we built two multiplayer clients with the same game logic, one using a first-person point of view, while the other includes a top-down camera perspective. Then, we conducted an online user study in which players tested these game clients in extensive multiplayer sessions. Analyzing speech time, in-game logs, questionnaires, and qualitative feedback, we look at the perspectives’ impact on player behavior, engagement, and game experience in a scientific or "serious games" context. In addition, we have made our designed game UNISON and both clients available as open source to facilitate future empirical social science research.
- Open Access
Relevant Irrelevance: Generating Alterfactual Explanations for Image Classifiers
In this paper, we demonstrate the feasibility of alterfactual explanations for black box image classifiers. Traditional explanation mechanisms from the field of Counterfactual Thinking are a widely-used paradigm for Explainable Artificial Intelligence (XAI), as they follow a natural way of reasoning that humans are familiar with. However, most common approaches from this field are based on communicating information about features or characteristics that are especially important for an AI's decision. However, to fully understand a decision, not only knowledge about relevant features is needed, but the awareness of irrelevant information also highly contributes to the creation of a user's mental model of an AI system. To this end, a novel approach for explaining AI systems called alterfactual explanations was recently proposed on a conceptual level. It is based on showing an alternative reality where irrelevant features of an AI's input are altered. By doing so, the user directly sees which input data characteristics can change arbitrarily without influencing the AI's decision. In this paper, we show for the first time that it is possible to apply this idea to black box models based on neural networks. To this end, we present a GAN-based approach to generate these alterfactual explanations for binary image classifiers. Further, we present a user study that gives interesting insights on how alterfactual explanations can complement counterfactual explanations.
Open AccessRelevant Irrelevance: Generating Alterfactual Explanations for Image Classifiers
In this paper, we demonstrate the feasibility of alterfactual explanations for black box image classifiers. Traditional explanation mechanisms from the field of Counterfactual Thinking are a widely-used paradigm for Explainable Artificial Intelligence (XAI), as they follow a natural way of reasoning that humans are familiar with. However, most common approaches from this field are based on communicating information about features or characteristics that are especially important for an AI's decision. However, to fully understand a decision, not only knowledge about relevant features is needed, but the awareness of irrelevant information also highly contributes to the creation of a user's mental model of an AI system. To this end, a novel approach for explaining AI systems called alterfactual explanations was recently proposed on a conceptual level. It is based on showing an alternative reality where irrelevant features of an AI's input are altered. By doing so, the user directly sees which input data characteristics can change arbitrarily without influencing the AI's decision. In this paper, we show for the first time that it is possible to apply this idea to black box models based on neural networks. To this end, we present a GAN-based approach to generate these alterfactual explanations for binary image classifiers. Further, we present a user study that gives interesting insights on how alterfactual explanations can complement counterfactual explanations.
- Open Access
- 18 March 2024
- Published by Association for Computing Machinery (ACM)
- p. 253-262
Does Difficulty even Matter? Investigating Difficulty Adjustment and Practice Behavior in an Open-Ended Learning Task
Difficulty adjustment in practice exercises has been shown to be beneficial for learning. However, previous research has mostly investigated close-ended tasks, which do not offer the students multiple ways to reach a valid solution. Contrary to this, in order to learn in an open-ended learning task, students need to effectively explore the solution space as there are multiple ways to reach a solution. For this reason, the effects of difficulty adjustment could be different for open-ended tasks. To investigate this, as our first contribution, we compare different methods of difficulty adjustment in a user study conducted with 86 participants. Furthermore, as the practice behavior of the students is expected to influence how well the students learn, we additionally look at their practice behavior as a post-hoc analysis. Therefore, as a second contribution, we identify different types of practice behavior and how they link to students’ learning outcomes and subjective evaluation measures as well as explore the influence the difficulty adjustment methods have on the practice behaviors. Our results suggest the usefulness of taking into account the practice behavior in addition to only using the practice performance to inform adaptive intervention and difficulty adjustment methods.
- 18 March 2024
- Published by Association for Computing Machinery (ACM)
- p. 253-262
Open AccessDoes Difficulty even Matter? Investigating Difficulty Adjustment and Practice Behavior in an Open-Ended Learning Task
Difficulty adjustment in practice exercises has been shown to be beneficial for learning. However, previous research has mostly investigated close-ended tasks, which do not offer the students multiple ways to reach a valid solution. Contrary to this, in order to learn in an open-ended learning task, students need to effectively explore the solution space as there are multiple ways to reach a solution. For this reason, the effects of difficulty adjustment could be different for open-ended tasks. To investigate this, as our first contribution, we compare different methods of difficulty adjustment in a user study conducted with 86 participants. Furthermore, as the practice behavior of the students is expected to influence how well the students learn, we additionally look at their practice behavior as a post-hoc analysis. Therefore, as a second contribution, we identify different types of practice behavior and how they link to students’ learning outcomes and subjective evaluation measures as well as explore the influence the difficulty adjustment methods have on the practice behaviors. Our results suggest the usefulness of taking into account the practice behavior in addition to only using the practice performance to inform adaptive intervention and difficulty adjustment methods.
- Open Access
Does Difficulty even Matter? Investigating Difficulty Adjustment and Practice Behavior in an Open-Ended Learning Task
Difficulty adjustment in practice exercises has been shown to be beneficial for learning. However, previous research has mostly investigated close-ended tasks, which do not offer the students multiple ways to reach a valid solution. Contrary to this, in order to learn in an open-ended learning task, students need to effectively explore the solution space as there are multiple ways to reach a solution. For this reason, the effects of difficulty adjustment could be different for open-ended tasks. To investigate this, as our first contribution, we compare different methods of difficulty adjustment in a user study conducted with 86 participants. Furthermore, as the practice behavior of the students is expected to influence how well the students learn, we additionally look at their practice behavior as a post-hoc analysis. Therefore, as a second contribution, we identify different types of practice behavior and how they link to students' learning outcomes and subjective evaluation measures as well as explore the influence the difficulty adjustment methods have on the practice behaviors. Our results suggest the usefulness of taking into account the practice behavior in addition to only using the practice performance to inform adaptive intervention and difficulty adjustment methods.
Open AccessDoes Difficulty even Matter? Investigating Difficulty Adjustment and Practice Behavior in an Open-Ended Learning Task
Difficulty adjustment in practice exercises has been shown to be beneficial for learning. However, previous research has mostly investigated close-ended tasks, which do not offer the students multiple ways to reach a valid solution. Contrary to this, in order to learn in an open-ended learning task, students need to effectively explore the solution space as there are multiple ways to reach a solution. For this reason, the effects of difficulty adjustment could be different for open-ended tasks. To investigate this, as our first contribution, we compare different methods of difficulty adjustment in a user study conducted with 86 participants. Furthermore, as the practice behavior of the students is expected to influence how well the students learn, we additionally look at their practice behavior as a post-hoc analysis. Therefore, as a second contribution, we identify different types of practice behavior and how they link to students' learning outcomes and subjective evaluation measures as well as explore the influence the difficulty adjustment methods have on the practice behaviors. Our results suggest the usefulness of taking into account the practice behavior in addition to only using the practice performance to inform adaptive intervention and difficulty adjustment methods.
- 12 July 2023
- Published by Springer Nature
- p. 320-332
Integrating Policy Summaries with Reward Decomposition for Explaining Reinforcement Learning Agents
- 12 July 2023
- Published by Springer Nature
- p. 320-332
Integrating Policy Summaries with Reward Decomposition for Explaining Reinforcement Learning Agents
- 9 July 2023
- Published by Springer Nature
- p. 381-397
Exploring the Effect of Visual-Based Subliminal Persuasion in Public Speeches Using Explainable AI Techniques
- 9 July 2023
- Published by Springer Nature
- p. 381-397
Exploring the Effect of Visual-Based Subliminal Persuasion in Public Speeches Using Explainable AI Techniques
- Open Access
GANterfactual-RL: Understanding Reinforcement Learning Agents' Strategies through Visual Counterfactual Explanations
Counterfactual explanations are a common tool to explain artificial intelligence models. For Reinforcement Learning (RL) agents, they answer "Why not?" or "What if?" questions by illustrating what minimal change to a state is needed such that an agent chooses a different action. Generating counterfactual explanations for RL agents with visual input is especially challenging because of their large state spaces and because their decisions are part of an overarching policy, which includes long-term decision-making. However, research focusing on counterfactual explanations, specifically for RL agents with visual input, is scarce and does not go beyond identifying defective agents. It is unclear whether counterfactual explanations are still helpful for more complex tasks like analyzing the learned strategies of different agents or choosing a fitting agent for a specific task. We propose a novel but simple method to generate counterfactual explanations for RL agents by formulating the problem as a domain transfer problem which allows the use of adversarial learning techniques like StarGAN. Our method is fully model-agnostic and we demonstrate that it outperforms the only previous method in several computational metrics. Furthermore, we show in a user study that our method performs best when analyzing which strategies different agents pursue.
Open AccessGANterfactual-RL: Understanding Reinforcement Learning Agents' Strategies through Visual Counterfactual Explanations
Counterfactual explanations are a common tool to explain artificial intelligence models. For Reinforcement Learning (RL) agents, they answer "Why not?" or "What if?" questions by illustrating what minimal change to a state is needed such that an agent chooses a different action. Generating counterfactual explanations for RL agents with visual input is especially challenging because of their large state spaces and because their decisions are part of an overarching policy, which includes long-term decision-making. However, research focusing on counterfactual explanations, specifically for RL agents with visual input, is scarce and does not go beyond identifying defective agents. It is unclear whether counterfactual explanations are still helpful for more complex tasks like analyzing the learned strategies of different agents or choosing a fitting agent for a specific task. We propose a novel but simple method to generate counterfactual explanations for RL agents by formulating the problem as a domain transfer problem which allows the use of adversarial learning techniques like StarGAN. Our method is fully model-agnostic and we demonstrate that it outperforms the only previous method in several computational metrics. Furthermore, we show in a user study that our method performs best when analyzing which strategies different agents pursue.
- Open Access
- 13 July 2022
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 5, 903875
Benchmarking Perturbation-Based Saliency Maps for Explaining Atari Agents
One of the most prominent methods for explaining the behavior of Deep Reinforcement Learning (DRL) agents is the generation of saliency maps that show how much each pixel attributed to the agents' decision. However, there is no work that computationally evaluates and compares the fidelity of different perturbation-based saliency map approaches specifically for DRL agents. It is particularly challenging to computationally evaluate saliency maps for DRL agents since their decisions are part of an overarching policy, which includes long-term decision making. For instance, the output neurons of value-based DRL algorithms encode both the value of the current state as well as the expected future reward after doing each action in this state. This ambiguity should be considered when evaluating saliency maps for such agents. In this paper, we compare five popular perturbation-based approaches to create saliency maps for DRL agents trained on four different Atari 2,600 games. The approaches are compared using two computational metrics: dependence on the learned parameters of the underlying deep Q-network of the agents (sanity checks) and fidelity to the agents' reasoning (input degradation). During the sanity checks, we found that a popular noise-based saliency map approach for DRL agents shows little dependence on the parameters of the output layer. We demonstrate that this can be fixed by tweaking the algorithm such that it focuses on specific actions instead of the general entropy within the output values. For fidelity, we identify two main factors that influence which saliency map approach should be chosen in which situation. Particular to value-based DRL agents, we show that analyzing the agents' choice of action requires different saliency map approaches than analyzing the agents' state value estimation.
- 13 July 2022
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 5, 903875
Open AccessBenchmarking Perturbation-Based Saliency Maps for Explaining Atari Agents
One of the most prominent methods for explaining the behavior of Deep Reinforcement Learning (DRL) agents is the generation of saliency maps that show how much each pixel attributed to the agents' decision. However, there is no work that computationally evaluates and compares the fidelity of different perturbation-based saliency map approaches specifically for DRL agents. It is particularly challenging to computationally evaluate saliency maps for DRL agents since their decisions are part of an overarching policy, which includes long-term decision making. For instance, the output neurons of value-based DRL algorithms encode both the value of the current state as well as the expected future reward after doing each action in this state. This ambiguity should be considered when evaluating saliency maps for such agents. In this paper, we compare five popular perturbation-based approaches to create saliency maps for DRL agents trained on four different Atari 2,600 games. The approaches are compared using two computational metrics: dependence on the learned parameters of the underlying deep Q-network of the agents (sanity checks) and fidelity to the agents' reasoning (input degradation). During the sanity checks, we found that a popular noise-based saliency map approach for DRL agents shows little dependence on the parameters of the output layer. We demonstrate that this can be fixed by tweaking the algorithm such that it focuses on specific actions instead of the general entropy within the output values. For fidelity, we identify two main factors that influence which saliency map approach should be chosen in which situation. Particular to value-based DRL agents, we show that analyzing the agents' choice of action requires different saliency map approaches than analyzing the agents' state value estimation.
- Open Access
- 1 December 2021
- Published by Elsevier in Artificial Intelligence
- Vol. 301
Local and global explanations of agent behavior: Integrating strategy summaries with saliency maps
- 1 December 2021
- Published by Elsevier in Artificial Intelligence
- Vol. 301
Open AccessLocal and global explanations of agent behavior: Integrating strategy summaries with saliency maps