4 results
Domain adaptation with transfer learning for pasture digital twins
- Christos Pylianidis, Michiel G.J. Kallenberg, Ioannis N. Athanasiadis
-
- Journal:
- Environmental Data Science / Volume 3 / 2024
- Published online by Cambridge University Press:
- 15 March 2024, e8
-
- Article
-
- You have access Access
- Open access
- HTML
- Export citation
-
Domain adaptation is important in agriculture because agricultural systems have their own individual characteristics. Applying the same treatment practices (e.g., fertilization) to different systems may not have the desired effect due to those characteristics. Domain adaptation is also an inherent aspect of digital twins. In this work, we examine the potential of transfer learning for domain adaptation in pasture digital twins. We use a synthetic dataset of grassland pasture simulations to pretrain and fine-tune machine learning metamodels for nitrogen response rate prediction. We investigate the outcome in locations with diverse climates, and examine the effect on the results of including more weather and agricultural management practices data during the pretraining phase. We find that transfer learning seems promising to make the models adapt to new conditions. Moreover, our experiments show that adding more weather data on the pretraining phase has a small effect on fine-tuned model performance compared to adding more management practices. This is an interesting finding that is worth further investigation in future studies.
Nitrogen management with reinforcement learning and crop growth models
- Michiel G.J. Kallenberg, Hiske Overweg, Ron van Bree, Ioannis N. Athanasiadis
-
- Journal:
- Environmental Data Science / Volume 2 / 2023
- Published online by Cambridge University Press:
- 25 September 2023, e34
-
- Article
-
- You have access Access
- Open access
- HTML
- Export citation
-
The growing need for agricultural products and the challenges posed by environmental and economic factors have created a demand for enhanced agricultural systems management. Machine learning has increasingly been leveraged to tackle agricultural optimization problems, and in particular, reinforcement learning (RL), a subfield of machine learning, seems a promising tool for data-driven discovery of future farm management policies. In this work, we present the development of CropGym, a Gymnasium environment, where a reinforcement learning agent can learn crop management policies using a variety of process-based crop growth models. As a use case, we report on the discovery of strategies for nitrogen application in winter wheat. An RL agent is trained to decide weekly on applying a discrete amount of nitrogen fertilizer, with the aim of achieving a balance between maximizing yield and minimizing environmental impact. Results show that close to optimal strategies are learned, competitive with standard practices set by domain experts. In addition, we evaluate, as an out-of-distribution test, whether the obtained policies are resilient against a change in climate conditions. We find that, when rainfall is sufficient, the RL agent remains close to the optimal policy. With CropGym, we aim to facilitate collaboration between the RL and agronomy communities to address the challenges of future agricultural decision-making.
Gender, age at onset, and duration of being ill as predictors for the long-term course and outcome of schizophrenia: an international multicenter study
- Konstantinos N. Fountoulakis, Elena Dragioti, Antonis T. Theofilidis, Tobias Wiklund, Xenofon Atmatzidis, Ioannis Nimatoudis, Erik Thys, Martien Wampers, Luchezar Hranov, Trayana Hristova, Daniil Aptalidis, Roumen Milev, Felicia Iftene, Filip Spaniel, Pavel Knytl, Petra Furstova, Tiina From, Henry Karlsson, Maija Walta, Raimo K. R. Salokangas, Jean-Michel Azorin, Justine Bouniard, Julie Montant, Georg Juckel, Ida S. Haussleiter, Athanasios Douzenis, Ioannis Michopoulos, Panagiotis Ferentinos, Nikolaos Smyrnis, Leonidas Mantonakis, Zsófia Nemes, Xenia Gonda, Dora Vajda, Anita Juhasz, Amresh Shrivastava, John Waddington, Maurizio Pompili, Anna Comparelli, Valentina Corigliano, Elmars Rancans, Alvydas Navickas, Jan Hilbig, Laurynas Bukelskis, Lidija I. Stevovic, Sanja Vodopic, Oluyomi Esan, Oluremi Oladele, Christopher Osunbote, Janusz K. Rybakowski, Pawel Wojciak, Klaudia Domowicz, Maria L. Figueira, Ludgero Linhares, Joana Crawford, Anca-Livia Panfil, Daria Smirnova, Olga Izmailova, Dusica Lecic-Tosevski, Henk Temmingh, Fleur Howells, Julio Bobes, Maria P. Garcia-Portilla, Leticia García-Alvarez, Gamze Erzin, Hasan Karadağ, Avinash De Sousa, Anuja Bendre, Cyril Hoschl, Cristina Bredicean, Ion Papava, Olivera Vukovic, Bojana Pejuskovic, Vincent Russell, Loukas Athanasiadis, Anastasia Konsta, Nikolaos K. Fountoulakis, Dan Stein, Michael Berk, Olivia Dean, Rajiv Tandon, Siegfried Kasper, Marc De Hert
-
- Journal:
- CNS Spectrums / Volume 27 / Issue 6 / December 2022
- Published online by Cambridge University Press:
- 09 August 2021, pp. 716-723
-
- Article
- Export citation
-
Background
The aim of the current study was to explore the effect of gender, age at onset, and duration on the long-term course of schizophrenia.
MethodsTwenty-nine centers from 25 countries representing all continents participated in the study that included 2358 patients aged 37.21 ± 11.87 years with a DSM-IV or DSM-5 diagnosis of schizophrenia; the Positive and Negative Syndrome Scale as well as relevant clinicodemographic data were gathered. Analysis of variance and analysis of covariance were used, and the methodology corrected for the presence of potentially confounding effects.
ResultsThere was a 3-year later age at onset for females (P < .001) and lower rates of negative symptoms (P < .01) and higher depression/anxiety measures (P < .05) at some stages. The age at onset manifested a distribution with a single peak for both genders with a tendency of patients with younger onset having slower advancement through illness stages (P = .001). No significant effects were found concerning duration of illness.
DiscussionOur results confirmed a later onset and a possibly more benign course and outcome in females. Age at onset manifested a single peak in both genders, and surprisingly, earlier onset was related to a slower progression of the illness. No effect of duration has been detected. These results are partially in accord with the literature, but they also differ as a consequence of the different starting point of our methodology (a novel staging model), which in our opinion precluded the impact of confounding effects. Future research should focus on the therapeutic policy and implications of these results in more representative samples.
Modeling psychological function in patients with schizophrenia with the PANSS: an international multi-center study
- Konstantinos N. Fountoulakis, Elena Dragioti, Antonis T. Theofilidis, Tobias Wiklund, Xenofon Atmatzidis, Ioannis Nimatoudis, Erik Thys, Martien Wampers, Luchezar Hranov, Trayana Hristova, Daniil Aptalidis, Roumen Milev, Felicia Iftene, Filip Spaniel, Pavel Knytl, Petra Furstova, Tiina From, Henry Karlsson, Maija Walta, Raimo K.R. Salokangas, Jean-Michel Azorin, Justine Bouniard, Julie Montant, Georg Juckel, Ida S. Haussleiter, Athanasios Douzenis, Ioannis Michopoulos, Panagiotis Ferentinos, Nikolaos Smyrnis, Leonidas Mantonakis, Zsófia Nemes, Xenia Gonda, Dora Vajda, Anita Juhasz, Amresh Shrivastava, John Waddington, Maurizio Pompili, Anna Comparelli, Valentina Corigliano, Elmars Rancans, Alvydas Navickas, Jan Hilbig, Laurynas Bukelskis, Lidija I. Stevovic, Sanja Vodopic, Oluyomi Esan, Oluremi Oladele, Christopher Osunbote, Janusz K. Rybakowski, Pawel Wojciak, Klaudia Domowicz, Maria L. Figueira, Ludgero Linhares, Joana Crawford, Anca-Livia Panfil, Daria Smirnova, Olga Izmailova, Dusica Lecic-Tosevski, Henk Temmingh, Fleur Howells, Julio Bobes, Maria P. Garcia-Portilla, Leticia García-Alvarez, Gamze Erzin, Hasan Karadağ, Avinash De Sousa, Anuja Bendre, Cyril Hoschl, Cristina Bredicean, Ion Papava, Olivera Vukovic, Bojana Pejuskovic, Vincent Russell, Loukas Athanasiadis, Anastasia Konsta, Dan Stein, Michael Berk, Olivia Dean, Rajiv Tandon, Siegfried Kasper, Marc De Hert
-
- Journal:
- CNS Spectrums / Volume 26 / Issue 3 / June 2021
- Published online by Cambridge University Press:
- 15 April 2020, pp. 290-298
-
- Article
- Export citation
-
Background
The aim of the current study was to explore the changing interrelationships among clinical variables through the stages of schizophrenia in order to assemble a comprehensive and meaningful disease model.
MethodsTwenty-nine centers from 25 countries participated and included 2358 patients aged 37.21 ± 11.87 years with schizophrenia. Multiple linear regression analysis and visual inspection of plots were performed.
ResultsThe results suggest that with progression stages, there are changing correlations among Positive and Negative Syndrome Scale factors at each stage and each factor correlates with all the others in that particular stage, in which this factor is dominant. This internal structure further supports the validity of an already proposed four stages model, with positive symptoms dominating the first stage, excitement/hostility the second, depression the third, and neurocognitive decline the last stage.
ConclusionsThe current study investigated the mental organization and functioning in patients with schizophrenia in relation to different stages of illness progression. It revealed two distinct “cores” of schizophrenia, the “Positive” and the “Negative,” while neurocognitive decline escalates during the later stages. Future research should focus on the therapeutic implications of such a model. Stopping the progress of the illness could demand to stop the succession of stages. This could be achieved not only by both halting the triggering effect of positive and negative symptoms, but also by stopping the sensitization effect on the neural pathways responsible for the development of hostility, excitement, anxiety, and depression as well as the deleterious effect on neural networks responsible for neurocognition.