EPEC - Engineering Predictability with Embodied Cognition (completed)

• Charles P. Martin, Kai Olav Ellefsen, and Jim Torresen. 2017. Deep models for ensemble touch-screen improvisation. Proceedings of Audio Mostly ’17.
• Charles P. Martin and Jim Torresen. 2017. Exploring social mobile music with tiny touch-screen performances. Proceedings of the 14th sound and music computing conference, Aalto University, pp. 175–180.
• Charles P. Martin and Jim Torresen. 2017. MicroJam: An app for sharing tiny touch-screen performances. Proceedings of the international conference on new interfaces for musical expression, Aalborg University Copenhagen, pp. 495–496.
• Nygaard, Tønnes Frostad; Tørresen, Jim; Glette, Kyrre Multi-objective Evolution of Fast and Stable Gaits on a Physical Quadruped Robotic Platform. Proceedings of the 2016 IEEE Symposium Series on Computational Intelligence (SSCI)
• Tørresen, Jim; Iversen, Andreas Høyer; Greisiger, Ralf Data from Past Patients used to Streamline Adjustment of Levels for Cochlear Implant for New Patients. Proceedings of the 2016 IEEE Symposium Series on Computational Intelligence (SSCI)

• Jørgen Nordmoen, Tønnes Frostad Nygaard, Eivind Samuelsen & Kyrre Glette (2021). On Restricting Real-Valued Genotypes in Evolutionary Algorithms. In Pedro A. Castillo & Juan Luiz Jiménez Laredo (Ed.), Applications of Evolutionary Computation. Springer. ISSN 9783030726997. p. 3–16. doi: 10.1007/978-3-030-72699-7_1. Full text in Research Archive Show summary

  Real-valued genotypes together with the variation operators, mutation and crossover, constitute some of the fundamental building blocks of Evolutionary Algorithms. Real-valued genotypes are utilized in a broad range of contexts, from weights in Artificial Neural Networks to parameters in robot control systems. Shared between most uses of real-valued genomes is the need for limiting the range of individual parameters to allowable bounds. In this paper we will illustrate the challenge of limiting the parameters of real-valued genomes and analyse the most promising method to properly limit these values. We utilize both empirical as well as benchmark examples to demonstrate the utility of the proposed method and through a literature review show how the insight of this paper could impact other research within the field. The proposed method requires minimal intervention from Evolutionary Algorithm practitioners and behaves well under repeated application of variation operators, leading to better theoretical properties as well as significant differences in well-known benchmarks.

• Tønnes Nygaard, Charles Patrick Martin, Jim Tørresen, Kyrre Glette & David Howard (2021). Real-world embodied AI through a morphologically adaptive quadruped robot. Nature Machine Intelligence. doi: 10.1038/s42256-021-00320-3. Full text in Research Archive Show summary

  Robots are traditionally bound by a fixed morphology during their operational lifetime, which is limited to adapting only their control strategies. Here we present the first quadrupedal robot that can morphologically adapt to different environmental conditions in outdoor, unstructured environments. Our solution is rooted in embodied AI and comprises two components: (1) a robot that permits in situ morphological adaptation and (2) an adaptation algorithm that transitions between the most energy-efficient morphologies on the basis of the currently sensed terrain. We first build a model that describes how the robot morphology affects performance on selected terrains. We then test continuous adaptation on realistic outdoor terrain while allowing the robot to constantly update its model. We show that the robot exploits its training to effectively transition between different morphological configurations, exhibiting substantial performance improvements over a non-adaptive approach. The demonstrated benefits of real-world morphological adaptation demonstrate the potential for a new embodied way of incorporating adaptation into future robotic designs.

• Tønnes Nygaard, Charles Patrick Martin, David Howard, Jim Tørresen & Kyrre Glette (2021). Environmental Adaptation of Robot Morphology and Control Through Real-world Evolution. Evolutionary Computation. ISSN 1063-6560. doi: 10.1162/evco_a_00291. Full text in Research Archive Show summary

  Robots operating in the real world will experience a range of different environments and tasks. It is essential for the robot to have the ability to adapt to its surroundings to work efficiently in changing conditions. Evolutionary robotics aims to solve this by optimizing both the control and body (morphology) of a robot, allowing adaptation to internal, as well as external factors. Most work in this field has been done in physics simulators, which are relatively simple and not able to replicate the richness of interactions found in the real world. Solutions that rely on the complex interplay between control, body, and environment are therefore rarely found. In this paper, we rely solely on real-world evaluations and apply evolutionary search to yield combinations of morphology and control for our mechanically self-reconfiguring quadruped robot. We evolve solutions on two distinct physical surfaces and analyze the results in terms of both control and morphology. We then transition to two previously unseen surfaces to demonstrate the generality of our method. We find that the evolutionary search finds high-performing and diverse morphology-controller configurations by adapting both control and body to the different properties of the physical environments. We additionally find that morphology and control vary with statistical significance between the environments. Moreover, we observe that our method allows for morphology and control parameters to transfer to previously-unseen terrains, demonstrating the generality of our approach.

• Tønnes Nygaard, David Howard & Kyrre Glette (2020). Real world morphological evolution is feasible. In Carlos A. Coello Coello (Eds.), GECCO '20: Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion. Association for Computing Machinery (ACM). ISSN 9781450371278. p. 1392–1394. doi: 10.1145/3377929.3398095. Full text in Research Archive Show summary

  Evolutionary algorithms offer great promise for the automatic design of robot bodies, tailoring them to specific environments or tasks. Most research is done on simplified models or virtual robots in physics simulators, which do not capture the natural noise and richness of the real world. Very few of these virtual robots are built as physical robots, and the few that are will rarely be further improved in the actual environment they operate in, limiting the effectiveness of the automatic design process. We utilize our shape-shifting quadruped robot, which allows us to optimize the design in its real-world environment. The robot is able to change the length of its legs during operation, and is robust enough for complex experiments and tasks. We have co-evolved control and morphology in several different scenarios, and have seen that the algorithm is able to exploit the dynamic morphology solely through real-world experiments.

• Charles Patrick Martin, Kyrre Glette, Tønnes Nygaard & Jim Tørresen (2020). Understanding Musical Predictions With an Embodied Interface for Musical Machine Learning. Frontiers in Artificial Intelligence. 3(6). doi: 10.3389/frai.2020.00006. Full text in Research Archive Show summary

  Machine-learning models of music often exist outside the worlds of musical performance practice and abstracted from the physical gestures of musicians. In this work, we consider how a recurrent neural network (RNN) model of simple music gestures may be integrated into a physical instrument so that predictions are sonically and physically entwined with the performer's actions. We introduce EMPI, an embodied musical prediction interface that simplifies musical interaction and prediction to just one dimension of continuous input and output. The predictive model is a mixture density RNN trained to estimate the performer's next physical input action and the time at which this will occur. Predictions are represented sonically through synthesized audio, and physically with a motorized output indicator. We use EMPI to investigate how performers understand and exploit different predictive models to make music through a controlled study of performances with different models and levels of physical feedback. We show that while performers often favor a model trained on human-sourced data, they find different musical affordances in models trained on synthetic, and even random, data. Physical representation of predictions seemed to affect the length of performances. This work contributes new understandings of how musicians use generative ML models in real-time performance backed up by experimental evidence. We argue that a constrained musical interface can expose the affordances of embodied predictive interactions.

• Andrei Faitas, Synne Engdahl Baumann, Torgrim Rudland Næss, Jim Tørresen & Charles Patrick Martin (2019). Generating Convincing Harmony Parts with Simple Long Short-Term Memory Networks. In Marcelo Queiroz & Anna Xambo Sedo (Ed.), Proceedings of the International Conference on New Interfaces for Musical Expression. Universidade Federal do Rio Grande do Sul. Full text in Research Archive Show summary

  Generating convincing music via deep neural networks is a challenging problem that shows promise for many applications including interactive musical creation. One part of this challenge is the problem of generating convincing accompaniment parts to a given melody, as could be used in an automatic accompaniment system. Despite much progress in this area, systems that can automatically learn to generate interesting and harmonically plausible accompaniments remain somewhat elusive. In this paper we explore systems where a user provides a sequence of notes, and a neural network model responds with an accompanying sequence of equal length. We consider two popular sequenceto- sequence models; one featuring standard unidirectional long short-term memory (LSTM) architecture, and the other featuring bidirectional LSTM. These are evaluated and compared via a qualitative study that features 106 respondents listening to eight random samples from our set of generated music, as well as two human samples. From the results we see a preference for the sequences generated by the bidirectional model as well as an indication that these sequences sound more human.

• Torgrim Rudland Næss & Charles Patrick Martin (2019). A Physical Intelligent Instrument using Recurrent Neural Networks. In Marcelo Queiroz & Anna Xambo Sedo (Ed.), Proceedings of the International Conference on New Interfaces for Musical Expression. Universidade Federal do Rio Grande do Sul. p. 79–82. Full text in Research Archive Show summary

  This paper describes a new intelligent interactive instrument, based on an embedded computing platform, where deep neural networks are applied to interactive music generation. Even though using neural networks for music composition is not uncommon, a lot of these models tend to not support any form of user interaction. We introduce a self-contained intelligent instrument using generative models, with support for real-time interaction where the user can adjust high-level parameters to modify the music generated by the instrument. We describe the technical details of our generative model and discuss the experience of using the system as part of musical performance.

• Charles Patrick Martin & Jim Tørresen (2019). An Interactive Musical Prediction System with Mixture Density Recurrent Neural Networks. In Marcelo Queiroz & Anna Xambo Sedo (Ed.), Proceedings of the International Conference on New Interfaces for Musical Expression. Universidade Federal do Rio Grande do Sul. p. 260–265. Full text in Research Archive Show summary

  This paper is about creating digital musical instruments where a predictive neural network model is integrated into the interactive system. Rather than predicting symbolic music (e.g., MIDI notes), we suggest that predicting future control data from the user and precise temporal information can lead to new and interesting interactive possibilities. We propose that a mixture density recurrent neural network (MDRNN) is an appropriate model for this task. The predictions can be used to fill-in control data when the user stops performing, or as a kind of filter on the user's own input. We present an interactive MDRNN prediction server that allows rapid prototyping of new NIMEs featuring predictive musical interaction by recording datasets, training MDRNN models, and experimenting with interaction modes. We illustrate our system with several example NIMEs applying this idea. Our evaluation shows that real-time predictive interaction is viable even on single-board computers and that small models are appropriate for small datasets.

• Tønnes Frostad Nygaard, Jørgen Halvorsen Nordmoen, Kai Olav Ellefsen, Charles Patrick Martin, Jim Tørresen & Kyrre Glette (2019). Experiences from Real-World Evolution with DyRET: Dynamic Robot for Embodied Testing, Nordic Artificial Intelligence Research and Development: Third Symposium of the Norwegian AI Society, NAIS 2019. Springer. ISSN 9783030356644. p. 58–68. doi: 10.1007/978-3-030-35664-4_6. Full text in Research Archive
• Kai Olav Ellefsen & Jim Tørresen (2019). Self-adapting Goals Allow Transfer of Predictive Models to New Tasks, Nordic Artificial Intelligence Research and Development: Third Symposium of the Norwegian AI Society, NAIS 2019. Springer. ISSN 9783030356644. p. 28–39. doi: 10.1007/978-3-030-35664-4_3. Full text in Research Archive
• Kai Olav Ellefsen, Joost Huizinga & Jim Tørresen (2019). Guiding Neuroevolution with Structural Objectives. Evolutionary Computation. ISSN 1063-6560. 28(1), p. 115–140. doi: 10.1162/evco_a_00250. Full text in Research Archive Show summary

  The structure and performance of neural networks are intimately connected, and by use of evolutionary algorithms, neural network structures optimally adapted to a given task can be explored. Guiding such neuroevolution with additional objectives related to network structure has been shown to improve performance in some cases, especially when modular neural networks are beneficial. However, apart from objectives aiming to make networks more modular, such structural objectives have not been widely explored. We propose two new structural objectives and test their ability to guide evolving neural networks on two problems which can benefit from decomposition into subtasks. The first structural objective guides evolution to align neural networks with a user-recommended decomposition pattern. Intuitively, this should be a powerful guiding target for problems where human users can easily identify a structure. The second structural objective guides evolution towards a population with a high diversity in decomposition patterns. This results in exploration of many different ways to decompose a problem, allowing evolution to find good decompositions faster. Tests on our target problems reveal that both methods perform well on a problem with a very clear and decomposable structure. However, on a problem where the optimal decomposition is less obvious, the structural diversity objective is found to outcompete other structural objectives—and this technique can even increase performance on problems without any decomposable structure at all.

• Justinas Miseikis, Inka Brijacak, Saeed Yahyanejad, Kyrre Glette, Ole Jacob Elle & Jim Tørresen (2019). Two-Stage Transfer Learning for Heterogeneous Robot Detection and 3D Joint Position Estimation in a 2D Camera Image Using CNN. IEEE International Conference on Robotics and Automation (ICRA). ISSN 1050-4729. 2019-May, p. 8883–8889. doi: 10.1109/ICRA.2019.8794077. Full text in Research Archive Show summary

  Collaborative robots are becoming more common on factory floors as well as regular environments, however, their safety still is not a fully solved issue. Collision detection does not always perform as expected and collision avoidance is still an active research area. Collision avoidance works well for fixed robot-camera setups, however, if they are shifted around, Eye-to-Hand calibration becomes invalid making it difficult to accurately run many of the existing collision avoidance algorithms. We approach the problem by presenting a stand-alone system capable of detecting the robot and estimating its position, including individual joints, by using a simple 2D colour image as an input, where no Eye-to-Hand calibration is needed. As an extension of previous work, a two-stage transfer learning approach is used to re-train a multi-objective convolutional neural network (CNN) to allow it to be used with heterogeneous robot arms. Our method is capable of detecting the robot in real-time and new robot types can be added by having significantly smaller training datasets compared to the requirements of a fully trained network. We present data collection approach, the structure of the multi-objective CNN, the two-stage transfer learning training and test results by using real robots from Universal Robots, Kuka, and Franka Emika. Eventually, we analyse possible application areas of our method together with the possible improvements.

• Tønnes Frostad Nygaard, Charles Patrick Martin, Jim Tørresen & Kyrre Glette (2019). Self-Modifying Morphology Experiments with DyRET: Dynamic Robot for Embodied Testing. IEEE International Conference on Robotics and Automation (ICRA). ISSN 1050-4729. 2019-May, p. 9446–9452. doi: 10.1109/ICRA.2019.8793663. Full text in Research Archive Show summary

  If robots are to become ubiquitous, they will need to be able to adapt to complex and dynamic environments. Robots that can adapt their bodies while deployed might be flexible and robust enough to meet this challenge. Previous work on dynamic robot morphology has focused on simulation, combining simple modules, or switching between locomotion modes. Here, we present an alternative approach: a self-reconfigurable morphology that allows a single four-legged robot to actively adapt the length of its legs to different environments. We report the design of our robot, as well as the results of a study that verifies the performance impact of self-reconfiguration. This study compares three different control and morphology pairs under different levels of servo supply voltage in the lab. We also performed preliminary tests in different uncontrolled outdoor environments to see if changes to the external environment supports our findings in the lab. Our results show better performance with an adaptable body, lending evidence to the value of self-reconfiguration for quadruped robots.

• Aline Weber, Lucas N. Alegre, Jim Tørresen & Bruno Castro da Silva (2019). Parameterized Melody Generation with Autoencoders and Temporally-Consistent Noise. In Federico Visi (Eds.), Music Proceedings of the International Conference on New Interfaces for Musical Expression. Universidade Federal do Rio Grande do Sul. p. 174–179. Full text in Research Archive Show summary

  We introduce a machine learning technique to autonomously generate novel melodies that are variations of an arbitrary base melody. These are produced by a neural network that ensures that (with high probability) the melodic and rhythmic structure of the new melody is consistent with a given set of sample songs. We train a Variational Autoencoder network to identify a low-dimensional set of variables that allows for the compression and representation of sample songs. By perturbing these variables with Perlin Noise— a temporally-consistent parameterized noise function—it is possible to generate smoothly-changing novel melodies. We show that (1) by regulating the amount of noise, one can specify how much of the base song will be preserved; and (2) there is a direct correlation between the noise signal and the differences between the statistical properties of novel melodies and the original one. Users can interpret the controllable noise as a type of “creativity knob”: the higher it is, the more leeway the network has to generate significantly different melodies. We present a physical prototype that allows musicians to use a keyboard to provide base melodies and to adjust the network’s “creativity knobs” to regulate in real-time the process that proposes new melody ideas.

• Benedikte Wallace & Charles Patrick Martin (2019). Comparing models for harmony prediction in an interactive audio looper. Lecture Notes in Computer Science (LNCS). ISSN 0302-9743. 11453 LNCS, p. 173–187. doi: 10.1007/978-3-030-16667-0_12. Full text in Research Archive
• Charles Patrick Martin & Henry Gardner (2019). Free-Improvised Rehearsal-as-Research for Musical HCI. In Simon Holland, Tom Mudd, Katie Wilkie-McKenna, Andrew McPherson & Marcelo M. Wanderley (Ed.), New Directions in Music and Human-Computer Interaction. Springer. ISSN 9783319920689. p. 269–284. doi: 10.1007/978-3-319-92069-6_17. Full text in Research Archive Show summary

  The formal evaluation of new interfaces for musical expression (NIMEs) in their use by ensembles of musicians is a challenging problem in human-computer interaction (HCI). NIMEs are designed to support creative expressions that are often improvised and unexpected. In the collaborative setting of a musical ensemble, interactions are complex and it can be almost impossible to directly evaluate the impact of interface variations. The evaluation environment also needs to be carefully considered. In the wild, concert pressures and practicalities limit experimental control. In the laboratory, studies may not sufficiently reflect real-world usage to make their conclusions relevant. To address some of these issues, we propose a methodology of rehearsal-as-research to study free improvisation by ensembles of NIME performers. In this methodology, evaluation sessions are structured to mirror established practices for improvisation training and performance development. Such sessions can allow controlled, order-balanced studies with extensive data collection in the style of factorial HCI experiments while preserving the artistic setting of a rehearsal. Experiment design, questionnaires, and objective measures such as session duration will be discussed along with two case studies.

• Tønnes Frostad Nygaard, Charles Patrick Martin, Jim Tørresen & Kyrre Glette (2019). Evolving Robots on Easy Mode: Towards a Variable Complexity Controller for Quadrupeds. Lecture Notes in Computer Science (LNCS). ISSN 0302-9743. 11454 LNCS, p. 616–632. doi: 10.1007/978-3-030-16692-2_41. Full text in Research Archive Show summary

  The complexity of a legged robot’s environment or task can inform how specialised its gait must be to ensure success. Evolving specialised robotic gaits demands many evaluations—acceptable for computer simulations, but not for physical robots. For some tasks, a more general gait, with lower optimization costs, could be satisfactory. In this paper, we introduce a new type of gait controller where complexity can be set by a single parameter, using a dynamic genotype-phenotype mapping. Low controller complexity leads to conservative gaits, while higher complexity allows more sophistication and high performance for demanding tasks, at the cost of optimization effort. We investigate the new controller on a virtual robot in simulations and do preliminary testing on a real-world robot. We show that having variable complexity allows us to adapt to different optimization budgets. With a high evaluation budget in simulation, a complex controller performs best. Moreover, real-world evolution with a limited evaluation budget indicates that a lower gait complexity is preferable for a relatively simple environment.

• Jørgen Halvorsen Nordmoen, Tønnes Frostad Nygaard, Kai Olav Ellefsen & Kyrre Glette (2019). Evolved embodied phase coordination enables robust quadruped robot locomotion. In Manuel López-Ibáñez (Eds.), GECCO '19: Proceedings of the Genetic and Evolutionary Computation Conference. Association for Computing Machinery (ACM). ISSN 9781450361118. p. 133–141. doi: 10.1145/3321707.3321762. Full text in Research Archive Show summary

  Overcoming robotics challenges in the real world requires resilient control systems capable of handling a multitude of environments and unforeseen events. Evolutionary optimization using simulations is a promising way to automatically design such control systems, however, if the disparity between simulation and the real world becomes too large, the optimization process may result in dysfunctional real-world behaviors. In this paper, we address this challenge by considering embodied phase coordination in the evolutionary optimization of a quadruped robot controller based on central pattern generators. With this method, leg phases, and indirectly also inter-leg coordination, are influenced by sensor feedback. By comparing two very similar control systems we gain insight into how the sensory feedback approach affects the evolved parameters of the control system, and how the performances differ in simulation, in transferal to the real world, and to different real-world environments. We show that evolution enables the design of a control system with embodied phase coordination which is more complex than previously seen approaches, and that this system is capable of controlling a real-world multi-jointed quadruped robot. The approach reduces the performance discrepancy between simulation and the real world, and displays robustness towards new environments.

• Bjørn Ivar Teigen, Kai Olav Ellefsen & Jim Tørresen (2019). A Categorization of Reinforcement Learning Exploration Techniques Which Facilitates Combination of Different Methods, Proceedings of the 9th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9781538681299. p. 189–194. doi: 10.1109/DEVLRN.2019.8850685. Full text in Research Archive
• Rafael Garcia, Alexandre Xavier Falcao, Alexandru C. Telea, Bruno Castro da Silva, Jim Tørresen & Joao Luiz Dihl Comba (2019). A Methodology for Neural Network Architectural Tuning Using Activation Occurrence Maps. In Chrisina Jayne & Zoltán Somogyvári (Ed.), 2019 International Joint Conference on Neural Networks (IJCNN). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9781728119854. doi: 10.1109/IJCNN.2019.8852223. Full text in Research Archive
• Charles Patrick Martin & Jim Tørresen (2019). Data Driven Analysis of Tiny Touchscreen Performance with MicroJam. Computer Music Journal. ISSN 0148-9267. 43(4). doi: 10.1162/comj_a_00536. Full text in Research Archive Show summary

  The widespread adoption of mobile devices, such as smartphones and tablets, has made touchscreens a common interface for musical performance. While new mobile music instrument have been investigated from design and user experience perspectives, there is little examination of the performers’ musical outputs. In this work, we introduce a constrained touchscreen performance app, MicroJam, designed to enable collaboration between performers, and engage in a data-driven analysis of more than 1600 performances using the app. MicroJam constrains performances to five seconds, and emphasises frequent and casual music making through a social media-inspired interface. Performers collaborate by replying to performances, adding new musical layers that are played back at the same time. Our analysis shows that users tend to focus on the centre and diagonals of the touchscreen area, and tend to swirl or swipe rather than tap. We also observe that while long swipes dominate the visual appearance of performances, the majority of interactions are short with limited expressive possibilities. Our findings enhance our understanding of how users perform in touchscreen apps and could be applied in future app designs for social musical interaction.

• Artur Becker, Henrik Herrebrøden, Victor Evaristo Gonzalez Sanchez, Kristian Nymoen, Carla Maria Dal Sasso Freitas & Jim Tørresen [Show all 7 contributors for this article] (2019). Functional Data Analysis of Rowing Technique Using Motion Capture Data. In Grisha Coleman (Eds.), Proceedings of the 6th International Conference on Movement and Computing. ACM Publications. ISSN 9781450376549. doi: 10.1145/3347122.3347135. Full text in Research Archive Show summary

  We present an approach to analyzing the motion capture data of rowers using bivariate functional principal component analysis (bfPCA). The method has been applied on data from six elite rowers rowing on an ergometer. The analyses of the upper and lower body coordination during the rowing cycle revealed significant differ- ences between the rowers, even though the data was normalized to account for differences in body dimensions. We make an argument for the use of bfPCA and other functional data analysis methods for the quantitative evaluation and description of technique in sports.

• Aline Weber, Charles Patrick Martin, Jim Tørresen & Bruno Castro da Silva (2019). Identifying Reusable Early-Life Options, Proceedings of the 9th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9781538681299. doi: 10.1109/DEVLRN.2019.8850725. Full text in Research Archive
• Justinas Miseikis, Inka Brijacak, Saeed Yahyanejad, Kyrre Glette, Ole Jacob Elle & Jim Tørresen (2018). Transfer Learning for Unseen Robot Detection and Joint Estimation on a Multi-Objective Convolutional Neural Network, 2018 IEEE International Conference on Intelligence and Safety for Robotics (ISR 2018). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9781538655481. p. 337–342. doi: 10.1109/IISR.2018.8535937. Full text in Research Archive Show summary

  A significant problem of using deep learning techniques is the limited amount of data available for training. There are some datasets available for the popular problems like item recognition and classification or self-driving cars, however, it is very limited for the industrial robotics field. In previous work, we have trained a multi-objective Convolutional Neural Network (CNN) to identify the robot body in the image and estimate 3D positions of the joints by using just a 2D image, but it was limited to a range of robots produced by Universal Robots (UR). In this work, we extend our method to work with a new robot arm - Kuka LBR iiwa, which has a significantly different appearance and an additional joint. However, instead of collecting large datasets once again, we collect a number of smaller datasets containing a few hundred frames each and use transfer learning techniques on the CNN trained on UR robots to adapt it to a new robot having different shapes and visual features. We have proven that transfer learning is not only applicable in this field, but it requires smaller well-prepared training datasets, trains significantly faster and reaches similar accuracy compared to the original method, even improving it on some aspects.

• Jørgen Halvorsen Nordmoen, Eivind Samuelsen, Kai Olav Ellefsen & Kyrre Glette (2018). Dynamic mutation in MAP-Elites for robotic repertoire generation, The 2018 Conference on Artificial Life. MIT Press. ISSN 9780262355766. p. 598–605. doi: 10.1162/isal_a_00110. Full text in Research Archive Show summary

  One of the core functions in most Evolutionary Algorithms is mutation. In complex search spaces, which are common in Evolutionary Robotics, mutation is often used both for optimizing existing solutions, described as exploitation, and for spanning the search space, called exploration. This presents a difficult challenge for researchers as mutation parameters must be selected with care in order to balance the two, often contradictory, effects. Strategies that vary mutation during the search often try to estimate these effects in order to modify the mutation parameters. In this regard MAP-Elites, a Quality Diversity algorithm, presents an interesting opportunity. Because factors related to exploration and exploitation are readily available during the search, optimization based on these factors could be utilized to improve the search. In this paper we study how online adaptation of mutation rate, dynamic mutation, affects MAP-Elites in order to gain insight into how the search process is affected by the mutation rate. Our study compares fixed and dynamic mutation parameters for two different complex gait controllers. The results show that dynamic mutation combines favorably with MAP-Elites and that there is a strong relation between mutation parameters and exploration that can be utilized.

• Jørgen Halvorsen Nordmoen, Kai Olav Ellefsen & Kyrre Glette (2018). Combining MAP-Elites and Incremental Evolution to Generate Gaits for a Mammalian Quadruped Robot. Lecture Notes in Computer Science (LNCS). ISSN 0302-9743. 10784 LNCS, p. 719–733. doi: 10.1007/978-3-319-77538-8_48. Full text in Research Archive Show summary

  Four-legged mammals are capable of showing a great variety of movement patterns, ranging from a simple walk to more complex movement such as trots and gallops. Imbuing this diversity to quadruped robots is of interest in order to improve both mobility and reach. Within the field of Evolutionary Robotics, Quality Diversity techniques have shown a remarkable ability to produce not only effective, but also highly diverse solutions. When applying this approach to four-legged robots an initial problem is to create viable movement patterns that do not fall. This difficulty stems from the challenging fitness gradient due to the mammalian morphology. In this paper we propose a solution to overcome this problem by implementing incremental evolution within the Quality Diversity framework. This allows us to evolve controllers that become more complex while at the same time utilizing the diversity produced by Quality Diversity. We show that our approach is able to generate high fitness solutions early in the search process, keep these solutions and perform a more open-ended search towards the end of evolution.

• Rafael Garcia, Alexandru C Telea, Bruno Castro da Silva, Jim Tørresen & Joao Luiz Dihl Comba (2018). A task-and-technique centered survey on visual analytics for deep learning model engineering. Computers & graphics. ISSN 0097-8493. 77, p. 30–49. doi: 10.1016/j.cag.2018.09.018. Full text in Research Archive
• Charles Patrick Martin & Jim Tørresen (2018). RoboJam: A musical mixture density network for collaborative touchscreen interaction. Lecture Notes in Computer Science (LNCS). ISSN 0302-9743. 10783 LNCS, p. 161–176. doi: 10.1007/978-3-319-77583-8_11. Full text in Research Archive Show summary

  RoboJam is a machine-learning system for generating music that assists users of a touchscreen music app by performing responses to their short improvisations. This system uses a recurrent artificial neural network to generate sequences of touchscreen interactions and absolute timings, rather than high-level musical notes. To accomplish this, RoboJam's network uses a mixture density layer to predict appropriate touch interaction locations in space and time. In this paper, we describe the design and implementation of RoboJam's network and how it has been integrated into a touchscreen music app. A preliminary evaluation analyses the system in terms of training, musical generation and user interaction.

• Justinas Miseikis, Patrick Knobelreiter, Inka Brijacak, Saeed Yahyanejad, Kyrre Glette & Ole Jacob Elle [Show all 7 contributors for this article] (2018). Robot localisation and 3D position estimation using a free-moving camera and cascaded convolutional neural networks, 2018 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM 2018). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9781538618554. p. 181–187. doi: 10.1109/AIM.2018.8452236. Full text in Research Archive Show summary

  Many works in collaborative robotics and human-robot interaction focuses on identifying and predicting human behaviour while considering the information about the robot itself as given. This can be the case when sensors and the robot are calibrated in relation to each other and often the reconfiguration of the system is not possible, or extra manual work is required. We present a deep learning based approach to remove the constraint of having the need for the robot and the vision sensor to be fixed and calibrated in relation to each other. The system learns the visual cues of the robot body and is able to localise it, as well as estimate the position of robot joints in 3D space by just using a 2D color image. The method uses a cascaded convolutional neural network, and we present the structure of the network, describe our own collected dataset, explain the network training and achieved results. A fully trained system shows promising results in providing an accurate mask of where the robot is located and a good estimate of its joints positions in 3D. The accuracy is not good enough for visual servoing applications yet, however, it can be sufficient for general safety and some collaborative tasks not requiring very high precision. The main benefit of our method is the possibility of the vision sensor to move freely. This allows it to be mounted on moving objects, for example, a body of the person or a mobile robot working in the same environment as the robots are operating in.

• Justinas Miseikis, Inka Brijacak, Saeed Yahyanejad, Kyrre Glette, Ole Jacob Elle & Jim Tørresen (2018). Multi-Objective Convolutional Neural Networks for Robot Localisation and 3D Position Estimation in 2D Camera Images, 2018 15th International Conference on Ubiquitous Robots (UR 2018). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9781538663356. p. 597–603. doi: 10.1109/URAI.2018.8441813. Full text in Research Archive Show summary

  The field of collaborative robotics and humanrobot interaction often focuses on the prediction of human behaviour, while assuming the information about the robot setup and configuration being known. This is often the case with fixed setups, which have all the sensors fixed and calibrated in relation to the rest of the system. However, it becomes a limiting factor when the system needs to be reconfigured or moved. We present a deep learning approach, which aims to solve this issue. Our method learns to identify and precisely localise the robot in 2D camera images, so having a fixed setup is no longer a requirement and a camera can be moved. In addition, our approach identifies the robot type and estimates the 3D position of the robot base in the camera image as well as 3D positions of each of the robot joints. Learning is done by using a multiobjective convolutional neural network with four previously mentioned objectives simultaneously using a combined loss function. The multi-objective approach makes the system more flexible and efficient by reusing some of the same features and diversifying for each objective in lower layers. A fully trained system shows promising results in providing an accurate mask of where the robot is located and an estimate of its base and joint positions in 3D. We compare the results to our previous approach of using cascaded convolutional neural networks.

View all works in NVA

• Kai Olav Ellefsen & Katharina J. Rohlfing (2019). Proceedings of the 9th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics. IEEE (Institute of Electrical and Electronics Engineers). ISBN 9781538681299. 340 p. Full text in Research Archive

View all works in NVA

• Kyrre Glette (2020). Evolutionary algorithms for intelligent robots. Full text in Research Archive
• Jørgen Halvorsen Nordmoen & Ingrid Fadelli (2019). A new method to enable robust locomotion in a quadruped robot. [Internet]. TechXplore. Full text in Research Archive
• Justinas Miseikis, Inka Brijacak, Saeed Yahyanejad, Kyrre Glette, Ole Jacob Elle & Jim Tørresen (2019). Two-Stage Transfer Learning for Heterogeneous Robot Detection and 3D Joint Position Estimation in a 2D Camera Image Using CNN. Full text in Research Archive
• Kai Olav Ellefsen, Joost Huizinga & Jim Tørresen (2019). Guiding Neuroevolution with Structural Objectives. Full text in Research Archive
• Tønnes Frostad Nygaard, Charles Patrick Martin, Jim Tørresen & Kyrre Glette (2019). Self-Modifying Morphology Experiments with DyRET: Dynamic Robot for Embodied Testing. Full text in Research Archive
• Tønnes Frostad Nygaard, Jørgen Halvorsen Nordmoen, Charles Patrick Martin, Jim Tørresen & Kyrre Glette (2019). Lessons Learned from Real-World Experiments with DyRET: the Dynamic Robot for Embodied Testing. Full text in Research Archive
• Kyrre Glette (2019). Kunstig intelligens for tilpasningsdyktige roboter. Full text in Research Archive
• Jim Tørresen (2019). Intelligent and Adaptive Robots in Real-World Environment. Full text in Research Archive Show summary

  240862 259293 247697 288285 262762

• Jim Tørresen (2019). Future and Ethical Perspectives of Robotics and AI. Full text in Research Archive
• Jim Tørresen (2019). Sensing Human State with Application in Older People Care and Mental Health Treatment. Full text in Research Archive
• Jun Miura & Jim Tørresen (2019). Intelligent Robot Technologies for Care and Lifestyle Support. Full text in Research Archive
• Joao Luiz Dihl Comba & Jim Tørresen (2019). Visual Data Analysis of Unstructured and Big Data. Full text in Research Archive
• Katharina J. Rohlfing & Jim Tørresen (2019). Explainability: an interactive view. Full text in Research Archive
• Jim Tørresen, Kyrre Glette & Kai Olav Ellefsen (2019). Intelligent, Adaptive Robots in Real-World Scenarios. Full text in Research Archive
• Jim Tørresen, Kyrre Glette & Kai Olav Ellefsen (2019). Adaptive Robot Body and Control for Real-World Environments. Full text in Research Archive
• Kai Olav Ellefsen (2019). Hva Kan Roboter Lære av Biologisk Liv? Full text in Research Archive
• Artur Becker, Henrik Herrebrøden, Victor Evaristo Gonzalez Sanchez, Kristian Nymoen, Carla Maria Dal Sasso Freitas & Jim Tørresen [Show all 7 contributors for this article] (2019). Functional Data Analysis of Rowing Technique Using Motion Capture Data. Full text in Research Archive Show summary

  We present an approach to analyzing the motion capture data ofrowers using bivariate functional principal component analysis(bfPCA). The method has been applied on data from six elite rowersrowing on an ergometer. The analyses of the upper and lower bodycoordination during the rowing cycle revealed significant differences between the rowers, even though the data was normalized toaccount for differences in body dimensions. We make an argumentfor the use of bfPCA and other functional data analysis methods forthe quantitative evaluation and description of technique in sports.

• Jim Tørresen (2019). Intelligent Robots and Systems in Real-World Environment. Full text in Research Archive
• Jim Tørresen (2019). Design and Control of Robots for Real-World Environment. Full text in Research Archive
• Tønnes Frostad Nygaard & Kyrre Glette (2019). Her er universitetets mest avanserte, selvlærende robot. [Journal]. Apollon. Full text in Research Archive
• Kai Olav Ellefsen & Jim Tørresen (2019). Evolutionary Robotics: Automatic design of robot bodies and control. Full text in Research Archive
• Jim Tørresen (2019). Supporting Older People with Robots for Independent Living. Full text in Research Archive Show summary

  247697 288285 262762

• Jim Tørresen (2019). Hva er kunstig intelligens? Full text in Research Archive
• Jim Tørresen (2019). Artificial Intelligence and Applications in Health and Care. Full text in Research Archive
• Jim Tørresen (2019). Kunstig intelligens – hvem, hva og hvor. (Eng. Artificial Intelligence – who, what and where). Full text in Research Archive
• Charles Patrick Martin & Jim Tørresen (2019). An Interactive Musical Prediction System with Mixture Density Recurrent Neural Networks. Full text in Research Archive
• Torgrim Rudland Næss, Jim Tørresen & Charles Patrick Martin (2019). A Physical Intelligent Instrument using Recurrent Neural Networks. Full text in Research Archive
• Andrei Faitas, Synne Engdahl Baumann, Jim Torresen & Charles Patrick Martin (2019). Generating Convincing Harmony Parts with Simple Long Short-Term Memory Networks. Full text in Research Archive
• Charles Patrick Martin & Jim Torresen (2019). An Interactive Music Prediction System with Mixture Density Recurrent Neural Networks. Full text in Research Archive
• Charles Patrick Martin, Torgrim Rudland Næss, Andrei Faitas & Synne Engdahl Baumann (2019). Session on Musical Prediction and Generation with Deep Learning. Full text in Research Archive
• Kyrre Glette, Tønnes Frostad Nygaard & Yngve Vogt (2019). Her er universitetets nest selvlærende robot. [Journal]. Teknisk ukeblad. Full text in Research Archive
• Kai Olav Ellefsen & Jim Tørresen (2019). Self-Adapting Goals Allow Transfer of Predictive Models to New Tasks. Full text in Research Archive
• Bjørn Ivar Teigen, Kai Olav Ellefsen & Jim Tørresen (2019). A Categorization of Reinforcement Learning Exploration Techniques Which Facilitates Combination of Different Methods. Full text in Research Archive
• Jørgen Halvorsen Nordmoen, Tønnes Frostad Nygaard, Kai Olav Ellefsen & Kyrre Glette (2019). Evolved embodied phase coordination enables robust quadruped robot locomotion. Full text in Research Archive
• Tønnes Frostad Nygaard, Jørgen Halvorsen Nordmoen, Kai Olav Ellefsen, Charles Patrick Martin, Jim Tørresen & Kyrre Glette (2019). Experiences from Real-World Evolution with DyRET: Dynamic Robot for Embodied Testing. Full text in Research Archive
• Jim Tørresen (2019). Making Robots Adaptive and Preferable to Humans. Full text in Research Archive
• Kyrre Glette (2019). Kunstig intelligens for tilpasningsdyktige roboter. Full text in Research Archive
• Jim Tørresen (2018). Intelligent Systems for Medical and Healthcare Applications. Full text in Research Archive
• Tønnes Frostad Nygaard & Luke Dormehl (2018). This robot taught itself how to walk and it’s as clumsy as a newborn deer. [Journal]. Digital Trends. Full text in Research Archive
• Tønnes Frostad Nygaard & Robbie Gonzales (2018). How a Flock of Drones Developed Collective Intelligence. [Journal]. Wired Science. Full text in Research Archive
• Tønnes Frostad Nygaard & Loukia Papadopoulos (2018). New Evolving Robot Teaches Itself to Walk Through Trial and Error. [Journal]. Interesting Engineering. Full text in Research Archive
• Jim Tørresen (2018). Remote Lab and Applications for High Performance and Embedded Architectures. Full text in Research Archive
• Tønnes Frostad Nygaard & Matt Simon (2018). The shape-shifting robot that evolves by falling down. [Journal]. Wired Science. Full text in Research Archive
• Hans Jonas Fossum Moen, Kyrre Glette, Tønnes Frostad Nygaard & Mette Johnsrud (2018). Fem felt der vi får en førerløs fremtid. [Internet]. Titan.uio.no. Full text in Research Archive
• Charles Patrick Martin (2018). Deep Predictive Models in Interactive Music. Full text in Research Archive
• Kyrre Glette (2018). Automatic design of bodies and behaviors for real-world robots. Full text in Research Archive
• Charles Patrick Martin, Kyrre Glette, Tønnes Frostad Nygaard & Jim Tørresen (2018). Self-Awareness in a Cyber-Physical Predictive Musical Interface. Full text in Research Archive Show summary

  We introduce a new self-contained and self-aware interface for musical expression where a recurrent neural network (RNN) is integrated into a physical instrument design. The system includes levers for physical input and output, a speaker system, and an integrated single-board computer. The RNN serves as an internal model of the user’s physical input, and predictions can replace or complement direct sonic and physical control by the user. We explore this device in terms of different interaction configurations and learned models according to frameworks of self-aware cyber-physical systems.

• Tønnes Frostad Nygaard, Charles Patrick Martin, Jim Tørresen & Kyrre Glette (2018). Exploring Mechanically Self-Reconfiguring Robots for Autonomous Design. Full text in Research Archive Show summary

  Evolutionary robotics has aimed to optimize robot control and morphology to produce better and more robust robots. Most previous research only addresses optimization of control, and does this only in simulation. We have developed a four-legged mammal-inspired robot that features a self-reconfiguring morphology. In this paper, we discuss the possibilities opened up by being able to efficiently do experiments on a changing morphology in the real world. We discuss present challenges for such a platform and potential experimental designs that could unlock new discoveries. Finally, we place our robot in its context within general developments in the field of evolutionary robotics, and consider what advances the future might hold.

• Charles Patrick Martin (2018). Predictive Music Systems for Interactive Performance. Full text in Research Archive Show summary

  Automatic music generation is a compelling task where much recent progress has been made with deep learning models. But how these models can be integrated into interactive music systems; how can they encourage or enhance the music making of human users? Musical performance requires prediction to operate instruments, and perform in groups. Predictive models can help interactive systems to understand their temporal context, and ensemble behaviour. Deep learning can allow data-driven models with a long memory of past states. This process could be termed "predictive musical interaction", where a predictive model is embedded in a musical interface, assisting users by predicting unknown states of musical processes. I’ll discuss a framework for predictive musical interaction including examples from our lab, and consider how this work could be applied more broadly in HCI and robotics. This talk will cover material from this paper: https://arxiv.org/abs/1801.10492

• Charles Patrick Martin, Kyrre Glette & Jim Tørresen (2018). Creative Prediction with Neural Networks. Full text in Research Archive Show summary

  The goal of this tutorial is to apply predictive machine learning models to creative data. The focus of the tutorial will be recurrent neural networks (RNNs), deep learning models that can be used to generate sequential and temporal data. RNNs can be applied to many kinds of creative data including text and music. They can learn the long-range structure from a corpus of data and “create” new sequences by predicting one element at a time. When embedded in a creative interface, they can be used for “predictive interaction” where a human collaborates with, influences, and is influenced by a generative neural network. We will walk through the fundamental steps for training creative RNNs using live-coded demonstrations with Python code in Jupyter Notebooks. These steps are: collecting and cleaning data, building and training an RNN, and developing predictive interactions. We will also have live demonstrations and interactive live-hacking of our creative RNN systems! You’re welcome to bring a laptop with python to the tutorial and load up our code examples, or to follow along with us on the screen!

• Enrique Alejandro Garcia Ceja, Kai Olav Ellefsen, Charles Patrick Martin & Jim Tørresen (2018). Prediction, Interaction, and User Behaviour. Full text in Research Archive Show summary

  The goal of this tutorial is to apply predictive machine learning models to human behaviour through a human computer interface. We will introduce participants to the key stages for developing predictive interaction in user-facing technologies: collecting and identifying data, applying machine learning models, and developing predictive interactions. Many of us are aware of recent advances in deep neural networks (DNNs) and other machine learning (ML) techniques; however, it is not always clear how we can apply these techniques in interactive and real-time applications. Apart from well-known examples such as image classification and speech recognition, what else can predictive ML models be used for? How can these computational intelligence techniques be deployed to help users? In this tutorial, we will show that ML models can be applied to many interactive applications to enhance users’ experience and engagement. We will demonstrate how sensor and user interaction data can be collected and investigated, modelled using classical ML and DNNs, and where predictions of these models can feed back into an interface. We will walk through these processes using live-coded demonstrations with Python code in Jupyter Notebooks so participants will be able to see our investigations live and take the example code home to apply in their own projects. Our demonstrations will be motivated from examples from our own research in creativity support tools, robotics, and modelling user behaviour. In creativity, we will show how streams of interaction data from a creative musical interface can be modelled with deep recurrent neural networks (RNNs). From this data, we can predict users’ future interactions, or the potential interactions of other users. This enables us to “fill in” parts of a tablet-based musical ensemble when other users are not available, or to continue a user’s composition with potential musical parts. In user behaviour, we will show how smartphone sensor data can be used to infer user contextual information such as physical activities. This contextual information can be used to trigger interactions in smart home or internet of things (IoT) environments, to help tune interactive applications to user’s needs, or to help track health data.

• Jim Tørresen (2018). Når etikk betyr alt. Full text in Research Archive
• Jim Tørresen (2018). Kunstig intelligens – hvem, hva og hvor. Full text in Research Archive
• Adrian Stoica & Jim Tørresen (2018). Robots on the Moon, and their Role in a Future Lunar Economy. Full text in Research Archive
• Jim Tørresen (2018). Ethical Robots and Autonomous Systems. Full text in Research Archive
• Tønnes Frostad Nygaard, Vegard Dønnem Søyseth, Jørgen Halvorsen Nordmoen & Kyrre Glette (2018). Stand with the DyRET robot. Full text in Research Archive
• Tønnes Frostad Nygaard (2018). Real-World Evolution Adapts Robot Morphology and Control to Hardware Limitations. Full text in Research Archive
• Charles Patrick Martin & Jim Tørresen (2018). Predictive Musical Interaction with MDRNNs. Full text in Research Archive
• Jim Tørresen (2018). Frelsende eller fatalt? [Journal]. Forskningsetikk. Full text in Research Archive
• Charles Patrick Martin (2018). Creative Prediction with Neural Networks. Full text in Research Archive
• Kai Olav Ellefsen (2018). Evolusjonær Robotikk: Automatisk design og kontroll av roboter. Full text in Research Archive
• Kai Olav Ellefsen & Jim Tørresen (2018). Evolutionary Robotics: Automatic design of robot controllers and bodies. Full text in Research Archive
• Vegard Dønnem Søyseth, Tønnes Frostad Nygaard, Charles Patrick Martin, Md Zia Uddin & Kai Olav Ellefsen (2018). ROBIN-Stand ved Cutting Edge 2018. Full text in Research Archive
• Jim Tørresen (2018). Artificial Intelligence Applied for Real-World Systems. Full text in Research Archive
• Jim Tørresen (2018). Artificial Intelligence – State-of-the-art. Full text in Research Archive
• Torgrim Rudland Næss, Charles Patrick Martin & Jim Tørresen (2019). A Physical Intelligent Instrument using Recurrent Neural Networks. Universitetet i Oslo. Full text in Research Archive Show summary

  Composing and playing music generally requires knowledge of music theory and exercise in instrument training. While traditional musical instruments often require years of arduous practice to master, intelligent musical systems can provide an easier introduction into music creation for novice users. This thesis describes the design and implementation of a novel intelligent instrument for interactive generation of music with recurrent neural networks, allowing users with little to no musical experience to explore musical ideas. Even though using neural networks for music composition is not a new concept, most previous work in this field does not ordinarily support user interaction, and is often dependent on general-purpose computers or expensive setups to implement. The proposed instrument is self-contained, running an RNN-based generative music model on a Raspberry Pi single-board computer for continuous generation of monophonic melodies that are sonified using a built-in speaker. It supports real-time interaction where the user can modify the generated music by adjusting a set of high-level parameters: sampling temperature (diversity), tempo, volume, instrument sound selection, and generative model selection. A user study with twelve participants was conducted to see the impact the different high-level parameter controls can have on a participant’s perceived feeling of control over the musical output from the instrument, and to evaluate the generative models trained on different datasets in terms of musical quality. The numerical ratings and open-ended answers were analyzed both quantitatively and qualitatively. The results show that the perceived feeling of control over the music was quite high, and the high-level parameter controls allowed participants to creatively engage with the instrument in the music-making process.

• Benedikte Wallace & Charles Patrick Martin (2018). Predictive songwriting with concatenative accompaniment. Universitetet i Oslo. Full text in Research Archive Show summary

  Musicians often use tools such as loop-pedals and multitrack recorders to assist in improvisation and songwriting. While these devices are useful in creating new compositions from scratch, they do not contribute to the composition directly. In recent years, new musical instruments, interfaces and controllers using machine learning algorithms to create new sounds, generate accompaniment or construct novel compositions, have become available for both professional musicians and novices to enjoy. This thesis describes the design, implementation and evaluation of a system for predictive songwriting and improvisation using concatenative accompaniment which has been given the nickname PSCA. In its most simple form, the PSCA functions as an audio looper for vocal improvisation, but the system also utilises machine learning approaches to predict suitable harmonies to accompany the playback loop. Two machine learning algorithms were compared and implemented into the PSCA to facilitate harmony prediction: the hidden Markov model (HMM) and the Bidirectional Long Short-Term Memory (BLSTM). The HMM and BLSTM algorithms are trained on a dataset of lead sheets in order to learn the relationship between the notes in a melody and the chord which accompanies it as well as learning dependencies between chords to model chord progressions. In quantitative testing, the BLSTM model was found to be able to learn the harmony prediction task more effectively than the HMM model, this was also supported by a qualitative analysis of musicians using the PSCA system. The system proposed in this thesis provides a novel approach in which these two machine learning models are compared with regards to prediction accuracy on the dataset as well as the perceived musicality of each model when used for harmony prediction in the PSCA. This approach results in a system which can contribute to the improvisation and songwriting process by adding harmonies to the audio loop on-the-fly.

• Jim Tørresen, Bjørn Ivar Teigen & Kai Olav Ellefsen (2018). An Active Learning Perspective on Exploration in Reinforcement Learning. Universitetet i Oslo. Full text in Research Archive Show summary

  Reinforcement learning is a research field concerned with automatically solving problems by trial and error. With each trial, a reinforcement learning agent extracts information about the problem. The agent can choose to apply its current knowledge and solve the problem as well as that information allows, or it can decide to do something new and untried with the hope that this will yield new and useful information. Balancing this trade-off is called the exploration vs. exploitation problem, and is the primary focus of this thesis. The work is organized around three related research questions. Exploration algorithms have proliferated in the last few years. There are differences in the way exploration techniques create exploratory behavior. The first part of this work presents a categorization of exploration methods based on these differences. Techniques from each category are then tested and analyzed to evaluate their benefits and weaknesses. The categorization scheme makes it possible to easily combine exploration techniques. The second part of this work shows how to combine different exploration methods. Experimental results demonstrate that such a combination can produce new exploration algorithms which significantly outperform methods from each category. The best combined method outperforms the best single-category method, and is more robust to changes in the other parameters of the learning system. Combining methods from different categories of exploration techniques gives better results, but there are still some relatively simple problems the combined methods can not efficiently solve. The third part of the work further investigates the family of exploration techniques known as intrinsic motivation. Inspired by how humans are motivated to explore, the field of intrinsic motivation creates curious AI agents. The current state of the art intrinsic motivation techniques work very well on many problems but can not solve problems where the solution requires modeling randomness. This work investigates a novel intrinsic motivation algorithm designed to remove this known weakness.

• Henrik Brustad & Charles Patrick Martin (2018). Digital Audio Generation with Neural Networks. Universitetet i Oslo. Full text in Research Archive

View all works in NVA