| 1. |
- Enrichi, Francesco, et al.
(författare)
-
Ag nanoaggregates as efficient broadband sensitizers for Tb3+ ions in silica-zirconia ion-exchanged sol-gel glasses and glass-ceramics
- 2018
-
Ingår i: Optical materials (Amsterdam). - : Elsevier. - 0925-3467 .- 1873-1252. ; 84, s. 668-674
-
Tidskriftsartikel (refereegranskat)abstract
- In this paper we report the study of down-shifting silica-zirconia glass and glass-ceramic films doped by Tb3+ ions and Ag nanoaggregates, which combine the typical spectral properties of the rare-earth-ions with the broadband sensitizing effect of the metal nanostructures. Na-Tb co-doped silica-zirconia samples were obtained by a modified sol-gel route. Dip-coating deposition followed by annealing for solvent evaporation and matrix densification were repeated several times, obtaining a homogeneous crack-free film. A final treatment at 700 °C or 1000 °C was performed to control the nanoscale structural properties of the samples, resulting respectively in a glass (G) or a glass-ceramic (GC), where tetragonal zirconia nanocrystals are surrounded by an amorphous silica matrix. Ag introduction was then achieved by ion-exchange in a molten salt bath, followed by annealing in air to control the migration and aggregation of the metal ions. The comparison of the structural, compositional and optical properties are presented for G and GC samples, providing evidence of highly efficient photoluminescence enhancement in both systems, slightly better in G than in GC samples, with a remarkable increase of the green Tb3+ PL emission at 330 nm excitation: 12 times for G and 8 times for GC samples. Furthermore, after Ag-exchange, the shape of Tb3+ excitation resembles the one of Ag ions/nanoaggregates, with a broad significant absorption in the whole UV-blue spectral region. This broadband enhanced downshifting could find potential applications in lighting devices and in PV solar cells.
|
|
| 2. |
- Strannegård, Claes, 1962, et al.
(författare)
-
The Ecosystem Path to General AI
- 2021
-
Tidskriftsartikel (övrigt vetenskapligt/konstnärligt)abstract
- We start by discussing the link between ecosystem simulators and general AI. Then we present the open-source ecosystem simulator Ecotwin, which is based on the game engine Unity and operates on ecosystems containing inanimate objects like mountains and lakes, as well as organisms such as animals and plants. Animal cognition is modeled by integrating three separate networks: (i) a reflex network for hard-wired reflexes; (ii) a happiness network that maps sensory data such as oxygen, water, energy, and smells, to a scalar happiness value; and (iii) a policy network for selecting actions. The policy network is trained with reinforcement learning (RL), where the reward signal is defined as the happiness difference from one time step to the next. All organisms are capable of either sexual or asexual reproduction, and they die if they run out of critical resources. We report results from three studies with Ecotwin, in which natural phenomena emerge in the models without being hardwired. First, we study a terrestrial ecosystem with wolves, deer, and grass, in which a Lotka-Volterra style population dynamics emerges. Second, we study a marine ecosystem with phytoplankton, copepods, and krill, in which a diel vertical migration behavior emerges. Third, we study an ecosystem involving lethal dangers, in which certain agents that combine RL with reflexes outperform pure RL agents.
|
|
| 3. |
- Strannegård, Claes, 1962, et al.
(författare)
-
The Ecosystem Path to AGI
- 2022
-
Ingår i: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). - Cham : Springer International Publishing. - 1611-3349 .- 0302-9743. ; 13154 LNAI, s. 269-278
-
Konferensbidrag (refereegranskat)abstract
- We start by discussing the link between ecosystem simulators and artificial general intelligence (AGI). Then we present the open-source ecosystem simulator Ecotwin, which is based on the game engine Unity and operates on ecosystems containing inanimate objects like mountains and lakes, as well as organisms, such as animals and plants. Animal cognition is modeled by integrating three separate networks: (i) a reflex network for hard-wired reflexes; (ii) a happiness network that maps sensory data such as oxygen, water, energy, and smells, to a scalar happiness value; and (iii) a policy network for selecting actions. The policy network is trained with reinforcement learning (RL), where the reward signal is defined as the happiness difference from one time step to the next. All organisms are capable of either sexual or asexual reproduction, and they die if they run out of critical resources. We report results from three studies with Ecotwin, in which natural phenomena emerge in the models without being hardwired. First, we study a terrestrial ecosystem with wolves, deer, and grass, in which a Lotka-Volterra style population dynamics emerges. Second, we study a marine ecosystem with phytoplankton, copepods, and krill, in which a diel vertical migration behavior emerges. Third, we study an ecosystem involving lethal dangers, in which certain agents that combine RL with reflexes outperform pure RL agents.
|
|
