Reiji Suzuki
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Reiji Suzuki has a total of 79 co-authors. Here are the top co-authors by number of shared publications:
- Takaya Arita
Graduate School of Informatics Nagoya University Furo-cho Chikusa-ku 464-8601 Nagoya Japan
42 shared publications Takaya Arita
Graduate School of Informatics, Nagoya University, Nagoya, Japan
18 shared publications- Hiroshi G. Okuno
Waseda University,Tokyo,Japan; Graduate School of Informatics,Kyoto University,Kyoto,Japan, Japan
8 shared publications - Kazuhiro Nakadai
Institute of Science Tokyo
5 shared publications Kazuhiro Nakadai
Department of Systems and Control Engineering, School of Engineering Tokyo Institute of Technology Tokyo Japan
3 shared publications- Michał Joachimczak
Graduate School of Information Science, Nagoya University, Nagoya, Japan
3 shared publications
Latest Publications
- Open Access
- 27 September 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (19) , 10481
Agent-Based Models of Sexual Selection in Bird Vocalizations Using Generative Approaches
The current agent-based evolutionary models for animal communication rely on simplified signal representations that differ significantly from natural vocalizations. We propose a novel agent-based evolutionary model based on text-to-audio (TTA) models to generate realistic animal vocalizations, advancing from VAE-based real-valued genotypes to TTA-based textual genotypes that generate bird songs using a fine-tuned Stable Audio Open 1.0 model. In our sexual selection framework, males vocalize songs encoded by their genotypes while females probabilistically select mates based on the similarity between males’ songs and their preference patterns, with mutations and crossovers applied to textual genotypes using a large language model (Gemma-3). As a proof of concept, we compared TTA-based and VAE-based sexual selection models for the Blue-and-white Flycatcher (Cyanoptila cyanomelana)’s songs and preferences. While the VAE-based model produces population clustering but constrains the evolution to a narrow region near the latent space’s origin where reconstructed songs remain clear, the TTA-based model enhances the genotypic and phenotypic diversity, drives song diversification, and fosters the creation of novel bird songs. Generated songs were validated by a virtual expert using the BirdNET classifier, confirming their acoustic realism through classification into related taxa. These findings highlight the potential of combining large language models and TTA models in agent-based evolutionary models for animal communication.
- 27 September 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (19) , 10481
Open AccessAgent-Based Models of Sexual Selection in Bird Vocalizations Using Generative Approaches
The current agent-based evolutionary models for animal communication rely on simplified signal representations that differ significantly from natural vocalizations. We propose a novel agent-based evolutionary model based on text-to-audio (TTA) models to generate realistic animal vocalizations, advancing from VAE-based real-valued genotypes to TTA-based textual genotypes that generate bird songs using a fine-tuned Stable Audio Open 1.0 model. In our sexual selection framework, males vocalize songs encoded by their genotypes while females probabilistically select mates based on the similarity between males’ songs and their preference patterns, with mutations and crossovers applied to textual genotypes using a large language model (Gemma-3). As a proof of concept, we compared TTA-based and VAE-based sexual selection models for the Blue-and-white Flycatcher (Cyanoptila cyanomelana)’s songs and preferences. While the VAE-based model produces population clustering but constrains the evolution to a narrow region near the latent space’s origin where reconstructed songs remain clear, the TTA-based model enhances the genotypic and phenotypic diversity, drives song diversification, and fosters the creation of novel bird songs. Generated songs were validated by a virtual expert using the BirdNET classifier, confirming their acoustic realism through classification into related taxa. These findings highlight the potential of combining large language models and TTA models in agent-based evolutionary models for animal communication.
- Open Access
Evolutionary ecology of words
We propose a model for the evolutionary ecology of words as one attempt to extend evolutionary game theory and agent-based models by utilizing the rich linguistic expressions of Large Language Models (LLMs). Our model enables the emergence and evolution of diverse and infinite options for interactions among agents. Within the population, each agent possesses a short word (or phrase) generated by an LLM and moves within a spatial environment. When agents become adjacent, the outcome of their interaction is determined by the LLM based on the relationship between their words, with the loser's word being replaced by the winner's. Word mutations, also based on LLM outputs, may occur. We conducted preliminary experiments assuming that ``strong animal species" would survive. The results showed that from an initial population consisting of well-known species, many species emerged both gradually and in a punctuated equilibrium manner. Each trial demonstrated the unique evolution of diverse populations, with one type of large species becoming dominant, such as terrestrial animals, marine life, or extinct species, which were ecologically specialized and adapted ones across diverse extreme habitats. We also conducted a long-term experiment with a large population, demonstrating the emergence and coexistence of diverse species.
Open AccessEvolutionary ecology of words
We propose a model for the evolutionary ecology of words as one attempt to extend evolutionary game theory and agent-based models by utilizing the rich linguistic expressions of Large Language Models (LLMs). Our model enables the emergence and evolution of diverse and infinite options for interactions among agents. Within the population, each agent possesses a short word (or phrase) generated by an LLM and moves within a spatial environment. When agents become adjacent, the outcome of their interaction is determined by the LLM based on the relationship between their words, with the loser's word being replaced by the winner's. Word mutations, also based on LLM outputs, may occur. We conducted preliminary experiments assuming that ``strong animal species" would survive. The results showed that from an initial population consisting of well-known species, many species emerged both gradually and in a punctuated equilibrium manner. Each trial demonstrated the unique evolution of diverse populations, with one type of large species becoming dominant, such as terrestrial animals, marine life, or extinct species, which were ecologically specialized and adapted ones across diverse extreme habitats. We also conducted a long-term experiment with a large population, demonstrating the emergence and coexistence of diverse species.
- Open Access
- 17 January 2025
- Published by MDPI AG in Electronics
- Vol. 14 (2) , 354
Eco-Evo-Devo in the Adaptive Evolution of Artificial Creatures Within a 3D Physical Environment
Eco-evo-devo is an interdisciplinary field integrating ecology, evolutionary biology, and developmental biology. Niche construction refers to the phenomenon where organisms alter selection pressures through ecological activities, and ecological inheritance is the consequence of niche construction, passed down through generations. Development encompasses the process by which an organism grows and changes its form and behavior throughout its lifetime. Despite emerging research on eco-evo-devo, the evolution of organisms involving complex physical structures remains underexplored. We aim to explore interactions between these components in the evolution of the adaptive morphology and behavior of artificial creatures. We constructed an evolutionary model where a population of rigid-bodied creatures must cross two valleys to reach a goal, using morphological development and object construction in a 3D physical environment. We found that the evolution of both lifetime development (LD) and niche construction (NC) showed complementary roles, with LD contributing to crossing one valley and NC the other. Regarding ecological inheritance (EI), we observed how NC adapts to the presence of inherited objects by placing fewer objects to construct structures. On the other hand, a large percentage of EI can cause obstacles to form, which hinders the population from moving towards the target and facilitates non-adaptive morphology, making the evolution unsuccessful.
- 17 January 2025
- Published by MDPI AG in Electronics
- Vol. 14 (2) , 354
Open AccessEco-Evo-Devo in the Adaptive Evolution of Artificial Creatures Within a 3D Physical Environment
Eco-evo-devo is an interdisciplinary field integrating ecology, evolutionary biology, and developmental biology. Niche construction refers to the phenomenon where organisms alter selection pressures through ecological activities, and ecological inheritance is the consequence of niche construction, passed down through generations. Development encompasses the process by which an organism grows and changes its form and behavior throughout its lifetime. Despite emerging research on eco-evo-devo, the evolution of organisms involving complex physical structures remains underexplored. We aim to explore interactions between these components in the evolution of the adaptive morphology and behavior of artificial creatures. We constructed an evolutionary model where a population of rigid-bodied creatures must cross two valleys to reach a goal, using morphological development and object construction in a 3D physical environment. We found that the evolution of both lifetime development (LD) and niche construction (NC) showed complementary roles, with LD contributing to crossing one valley and NC the other. Regarding ecological inheritance (EI), we observed how NC adapts to the presence of inherited objects by placing fewer objects to construct structures. On the other hand, a large percentage of EI can cause obstacles to form, which hinders the population from moving towards the target and facilitates non-adaptive morphology, making the evolution unsuccessful.
- Open Access
- 5 August 2024
- Published by Wiley in Methods in Ecology and Evolution
- Vol. 15 (9) , 1568-1579
Spatial ecosystem monitoring with a Multichannel Acoustic Autonomous Recording Unit (MAARU)
Multi‐microphone recording adds spatial information to recorded audio with emerging applications in ecosystem monitoring. Specifically placing sounds in space can improve animal count accuracy, locate illegal activity like logging and poaching, track animals to monitor behaviour and habitat use and allow for ‘beamforming’ to amplify sounds from target directions for downstream classification. Studies have shown many advantages of spatial acoustics, but uptake remains limited as the equipment is often expensive, complicated, inaccessible or only suitable for short‐term deployments. With an emphasis on enhanced uptake and usability, we present a low‐cost, open‐source, six‐channel recorder built entirely from commercially available components which can be integrated into a solar‐powered, online system. The MAARU (Multichannel Acoustic Autonomous Recording Unit) works as an independent node in long‐term autonomous, passive and/or short‐term deployments. Here, we introduce MAARU's hardware and software and present the results of lab and field tests investigating the device's durability and usability. MAARU records multichannel audio with similar costs and power demands to equivalent omnidirectional recorders. MAARU devices have been deployed in the United Kingdom and Brazil, where we have shown MAARUs can accurately localise pure tones up to 6 kHz and bird calls as far as 8 m away (±10° range, 100% and >60% of signals, respectively). Louder calls may have even further detection radii. We also show how beamforming can be used with MAARUs to improve species ID confidence scores. MAARU is an accessible, low‐cost option for those looking to explore spatial acoustics accurately and easily with a single device, and without the formidable expenses and processing complications associated with establishing arrays. Ultimately, the added directional element of the multichannel recording provided by MAARU allows for enhanced recording of sonic environments, further enabling a potential step change in the uptake of spatial acoustics in the wider field.
- 5 August 2024
- Published by Wiley in Methods in Ecology and Evolution
- Vol. 15 (9) , 1568-1579
Open AccessSpatial ecosystem monitoring with a Multichannel Acoustic Autonomous Recording Unit (MAARU)
Multi‐microphone recording adds spatial information to recorded audio with emerging applications in ecosystem monitoring. Specifically placing sounds in space can improve animal count accuracy, locate illegal activity like logging and poaching, track animals to monitor behaviour and habitat use and allow for ‘beamforming’ to amplify sounds from target directions for downstream classification. Studies have shown many advantages of spatial acoustics, but uptake remains limited as the equipment is often expensive, complicated, inaccessible or only suitable for short‐term deployments. With an emphasis on enhanced uptake and usability, we present a low‐cost, open‐source, six‐channel recorder built entirely from commercially available components which can be integrated into a solar‐powered, online system. The MAARU (Multichannel Acoustic Autonomous Recording Unit) works as an independent node in long‐term autonomous, passive and/or short‐term deployments. Here, we introduce MAARU's hardware and software and present the results of lab and field tests investigating the device's durability and usability. MAARU records multichannel audio with similar costs and power demands to equivalent omnidirectional recorders. MAARU devices have been deployed in the United Kingdom and Brazil, where we have shown MAARUs can accurately localise pure tones up to 6 kHz and bird calls as far as 8 m away (±10° range, 100% and >60% of signals, respectively). Louder calls may have even further detection radii. We also show how beamforming can be used with MAARUs to improve species ID confidence scores. MAARU is an accessible, low‐cost option for those looking to explore spatial acoustics accurately and easily with a single device, and without the formidable expenses and processing complications associated with establishing arrays. Ultimately, the added directional element of the multichannel recording provided by MAARU allows for enhanced recording of sonic environments, further enabling a potential step change in the uptake of spatial acoustics in the wider field.
- Top Cited PapersOpen Access
- 19 March 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-9
An evolutionary model of personality traits related to cooperative behavior using a large language model
This study aims to demonstrate that Large Language Models (LLMs) can empower research on the evolution of human behavior, based on evolutionary game theory, by using an evolutionary model positing that instructing LLMs with high-level psychological and cognitive character descriptions enables the simulation of human behavior choices in game-theoretical scenarios. As a first step towards this objective, this paper proposes an evolutionary model of personality traits related to cooperative behavior using a large language model. In the model, linguistic descriptions of personality traits related to cooperative behavior are used as genes. The deterministic strategies extracted from LLM that make behavioral decisions based on these personality traits are used as behavioral traits. The population is evolved according to selection based on average payoff and mutation of genes by asking LLM to slightly modify the parent gene toward cooperative or selfish. Through experiments and analyses, we clarify that such a model can indeed exhibit evolution of cooperative behavior based on the diverse and higher-order representation of personality traits. We also observed repeated intrusion of cooperative and selfish personality traits through changes in the expression of personality traits. The words that emerged in the evolved genes reflected the behavioral tendencies of their associated personalities in terms of semantics, thereby influencing individual behavior and, consequently, the evolutionary dynamics.
- 19 March 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-9
Top Cited PapersOpen AccessAn evolutionary model of personality traits related to cooperative behavior using a large language model
This study aims to demonstrate that Large Language Models (LLMs) can empower research on the evolution of human behavior, based on evolutionary game theory, by using an evolutionary model positing that instructing LLMs with high-level psychological and cognitive character descriptions enables the simulation of human behavior choices in game-theoretical scenarios. As a first step towards this objective, this paper proposes an evolutionary model of personality traits related to cooperative behavior using a large language model. In the model, linguistic descriptions of personality traits related to cooperative behavior are used as genes. The deterministic strategies extracted from LLM that make behavioral decisions based on these personality traits are used as behavioral traits. The population is evolved according to selection based on average payoff and mutation of genes by asking LLM to slightly modify the parent gene toward cooperative or selfish. Through experiments and analyses, we clarify that such a model can indeed exhibit evolution of cooperative behavior based on the diverse and higher-order representation of personality traits. We also observed repeated intrusion of cooperative and selfish personality traits through changes in the expression of personality traits. The words that emerged in the evolved genes reflected the behavioral tendencies of their associated personalities in terms of semantics, thereby influencing individual behavior and, consequently, the evolutionary dynamics.
- Open Access
- 7 December 2023
- Vol. 7 (1)
Visualization and Quantification of the Activities of Animal Vocalizations in Forest Species Using Robot Audition Techniques
Ecoacoustics is a recent ecological discipline focusing on the ecological role of sounds. This study applies robot audition techniques and ecoacoustic methods in the visualization and quantification of forest animal vocalization, focusing mainly on inter-specific interactions between birds and cicadas. We adopt HARK, an open-source robot audition software, which enables us to estimate the direction of arrival of sound sources and their separated sounds using a microphone array unit. We focus on recordings in an experimental forest in Japan, where birds and cicadas dominate the soundscape. Cicada songs were further replayed at a regular interval repeatedly. We create a false-colour spectrogram based on the directions of arrival of sounds to grasp the individual-level dynamics of the soundscape of birds and cicadas in the recording. We further describe a method to classify their vocalizations using three ecoacoustic indices, then illustrate their temporal vocalization dynamics, measured as the total song duration in each time segment. We also conducted a quantitative analysis of their vocal activities to determine if there exist interactions among birds and cicadas, and the effects of replayed vocalizations of cicadas on them. The preliminary analysis implied that there might exist temporal overlap avoidance behaviours between birds and cicadas, and replayed songs of cicadas may reduce the activity of birds. We believe that this proof-of-concept observation and analysis can contribute to the further development of the fine-scale measurement of the biodiversity or habitat quality of environments based on the vocal activities of multiple species, while further detailed analysis is necessary.
- 7 December 2023
- Vol. 7 (1)
Open AccessVisualization and Quantification of the Activities of Animal Vocalizations in Forest Species Using Robot Audition Techniques
Ecoacoustics is a recent ecological discipline focusing on the ecological role of sounds. This study applies robot audition techniques and ecoacoustic methods in the visualization and quantification of forest animal vocalization, focusing mainly on inter-specific interactions between birds and cicadas. We adopt HARK, an open-source robot audition software, which enables us to estimate the direction of arrival of sound sources and their separated sounds using a microphone array unit. We focus on recordings in an experimental forest in Japan, where birds and cicadas dominate the soundscape. Cicada songs were further replayed at a regular interval repeatedly. We create a false-colour spectrogram based on the directions of arrival of sounds to grasp the individual-level dynamics of the soundscape of birds and cicadas in the recording. We further describe a method to classify their vocalizations using three ecoacoustic indices, then illustrate their temporal vocalization dynamics, measured as the total song duration in each time segment. We also conducted a quantitative analysis of their vocal activities to determine if there exist interactions among birds and cicadas, and the effects of replayed vocalizations of cicadas on them. The preliminary analysis implied that there might exist temporal overlap avoidance behaviours between birds and cicadas, and replayed songs of cicadas may reduce the activity of birds. We believe that this proof-of-concept observation and analysis can contribute to the further development of the fine-scale measurement of the biodiversity or habitat quality of environments based on the vocal activities of multiple species, while further detailed analysis is necessary.
- Open Access
An evolutionary model of personality traits related to cooperative behavior using a large language model
This paper aims to shed light on the evolutionary dynamics of diverse and social populations by introducing the rich expressiveness of generative models into the trait expression of social agent-based evolutionary models. Specifically, we focus on the evolution of personality traits in the context of a game-theoretic relationship as a situation in which inter-individual interests exert strong selection pressures. We construct an agent model in which linguistic descriptions of personality traits related to cooperative behavior are used as genes. The deterministic strategies extracted from Large Language Model (LLM) that make behavioral decisions based on these personality traits are used as behavioral traits. The population is evolved according to selection based on average payoff and mutation of genes by asking LLM to slightly modify the parent gene toward cooperative or selfish. Through preliminary experiments and analyses, we clarify that such a model can indeed exhibit the evolution of cooperative behavior based on the diverse and higher-order representation of personality traits. We also observed the repeated intrusion of cooperative and selfish personality traits through changes in the expression of personality traits, and found that the emerging words in the evolved gene well reflected the behavioral tendency of its personality in terms of their semantics.
Open AccessAn evolutionary model of personality traits related to cooperative behavior using a large language model
This paper aims to shed light on the evolutionary dynamics of diverse and social populations by introducing the rich expressiveness of generative models into the trait expression of social agent-based evolutionary models. Specifically, we focus on the evolution of personality traits in the context of a game-theoretic relationship as a situation in which inter-individual interests exert strong selection pressures. We construct an agent model in which linguistic descriptions of personality traits related to cooperative behavior are used as genes. The deterministic strategies extracted from Large Language Model (LLM) that make behavioral decisions based on these personality traits are used as behavioral traits. The population is evolved according to selection based on average payoff and mutation of genes by asking LLM to slightly modify the parent gene toward cooperative or selfish. Through preliminary experiments and analyses, we clarify that such a model can indeed exhibit the evolution of cooperative behavior based on the diverse and higher-order representation of personality traits. We also observed the repeated intrusion of cooperative and selfish personality traits through changes in the expression of personality traits, and found that the emerging words in the evolved gene well reflected the behavioral tendency of its personality in terms of their semantics.
- Open Access
- 31 March 2023
- Published by Wiley in Ecology and Evolution
- Vol. 13 (4) , e9938
Monitoring the courtship flight trajectory of Latham's snipe (Gallinago hardwickii) using microphone arrays
This study is the first to quantitatively measure the courtship display flights of Latham's snipe (Gallinago hardwickii), which is a “near threatened” species as of 2022 (IUCN red list of threatened species). By using a 16‐channel microphone array and 8‐channel microphone arrays, we localized the fine‐scale movements of courtship flights of one male performing at high altitude and high speed, and we estimated the direction from which each sound arrived using robot audition. Preliminary analyses of the azimuthal and elevation angles of the courtship flights partially revealed a fine‐scale flight trajectory. First, a male Latham's snipe gradually gained altitude while vocalizing sharp and harsh repeating calls, until it reached the flight peak altitude, then dove down while producing winnowing sound to the ground along the wetland zones without tall vegetation. This observation method is methodologically useful to establish a better understanding of Latham's snipe courtship flight site selection. Furthermore, this method can be extended to investigate other rare nocturnal or crepuscular birds that are too timid to risk ringing or tagging.
- 31 March 2023
- Published by Wiley in Ecology and Evolution
- Vol. 13 (4) , e9938
Open AccessMonitoring the courtship flight trajectory of Latham's snipe (Gallinago hardwickii) using microphone arrays
This study is the first to quantitatively measure the courtship display flights of Latham's snipe (Gallinago hardwickii), which is a “near threatened” species as of 2022 (IUCN red list of threatened species). By using a 16‐channel microphone array and 8‐channel microphone arrays, we localized the fine‐scale movements of courtship flights of one male performing at high altitude and high speed, and we estimated the direction from which each sound arrived using robot audition. Preliminary analyses of the azimuthal and elevation angles of the courtship flights partially revealed a fine‐scale flight trajectory. First, a male Latham's snipe gradually gained altitude while vocalizing sharp and harsh repeating calls, until it reached the flight peak altitude, then dove down while producing winnowing sound to the ground along the wetland zones without tall vegetation. This observation method is methodologically useful to establish a better understanding of Latham's snipe courtship flight site selection. Furthermore, this method can be extended to investigate other rare nocturnal or crepuscular birds that are too timid to risk ringing or tagging.
- Open Access
- 11 March 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (6) , 3607
Estimating the Soundscape Structure and Dynamics of Forest Bird Vocalizations in an Azimuth-Elevation Space Using a Microphone Array
Songbirds are one of the study targets for both bioacoustic and ecoacoustic research. In this paper, we discuss the applicability of robot audition techniques to understand the dynamics of forest bird vocalizations in a soundscape measured in azimuth and elevation angles with a single 16-channel microphone array, using HARK and HARKBird. First, we evaluated the accuracy in estimating the azimuth and elevation angles of bird vocalizations replayed from a loudspeaker on a tree, 6.55 m above the height of the array, from different horizontal distances in a forest. The results showed that the localization error of azimuth and elevation angle was equal to or less than 5 degrees and 15 degrees, respectively, in most of cases when the horizontal distance from the array was equal to or less than 35 m. We then conducted a field observation of vocalizations to monitor birds in a forest. The results showed that the system can successfully detect how birds use the soundscape horizontally and vertically. This can contribute to bioacoustic and ecoacoustic research, including behavioral observations and study of biodiversity.
- 11 March 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (6) , 3607
Open AccessEstimating the Soundscape Structure and Dynamics of Forest Bird Vocalizations in an Azimuth-Elevation Space Using a Microphone Array
Songbirds are one of the study targets for both bioacoustic and ecoacoustic research. In this paper, we discuss the applicability of robot audition techniques to understand the dynamics of forest bird vocalizations in a soundscape measured in azimuth and elevation angles with a single 16-channel microphone array, using HARK and HARKBird. First, we evaluated the accuracy in estimating the azimuth and elevation angles of bird vocalizations replayed from a loudspeaker on a tree, 6.55 m above the height of the array, from different horizontal distances in a forest. The results showed that the localization error of azimuth and elevation angle was equal to or less than 5 degrees and 15 degrees, respectively, in most of cases when the horizontal distance from the array was equal to or less than 35 m. We then conducted a field observation of vocalizations to monitor birds in a forest. The results showed that the system can successfully detect how birds use the soundscape horizontally and vertically. This can contribute to bioacoustic and ecoacoustic research, including behavioral observations and study of biodiversity.
- 17 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Extracting Bird Vocalizations from a Complex Natural Soundscape in Forests Using Robot Audition Techniques
Songbirds are important targets of bioacoustic and ecoacoustic research. We propose a simple method to extract bird vocalizations from many localized and separated sounds using robot audition techniques and report its evaluation by applying it to field recordings of birdsongs. The method utilizes latent space representations of the spectrogram based on a variational autoencoder and a distance measure to extract and classify separated songs. A preliminary analysis of field recording in California successfully extracted bird vocalizations, including ones from a considerably distant bird. We also illustrate an example of applying the method to a 2D localization of birdsongs during a playback experiment.
- 17 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Extracting Bird Vocalizations from a Complex Natural Soundscape in Forests Using Robot Audition Techniques
Songbirds are important targets of bioacoustic and ecoacoustic research. We propose a simple method to extract bird vocalizations from many localized and separated sounds using robot audition techniques and report its evaluation by applying it to field recordings of birdsongs. The method utilizes latent space representations of the spectrogram based on a variational autoencoder and a distance measure to extract and classify separated songs. A preliminary analysis of field recording in California successfully extracted bird vocalizations, including ones from a considerably distant bird. We also illustrate an example of applying the method to a 2D localization of birdsongs during a playback experiment.