• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.8
• /
• pp.331-338
• /
• 2022

Successor representation (SR) is a model of human reinforcement learning (RL) mimicking the underlying mechanism of hippocampal cells constructing cognitive maps. SR utilizes these learned features to adaptively respond to the frequent reward changes. In this paper, we evaluated the performance of SR under the context where changes in latent variables of environments trigger the reward structure changes. For a benchmark test, we adopted SR-Dyna, an integration of SR into goal-driven Dyna RL algorithm in the 2-stage Markov Decision Task (MDT) in which we can intentionally manipulate the latent variables - state transition uncertainty and goal-condition. To precisely investigate the characteristics of SR, we conducted the experiments while controlling each latent variable that affects the changes in reward structure. Evaluation results showed that SR-Dyna could learn to respond to the reward changes in relation to the changes in latent variables, but could not learn rapidly in that situation. This brings about the necessity to build more robust RL models that can rapidly learn to respond to the frequent changes in the environment in which latent variables and reward structure change at the same time.

• Korean Journal of Poultry Science
• /
• v.28 no.2
• /
• pp.83-89
• /
• 2001

Newcastle disease virus, CBP-1 strain isolated from Korean pheasants was passaged for 173 times by 9-day-old specific pathogenic free (SPF) embryonated eggs at $37^{\circ}C$ (parent strain) and subsequently passaged for 15 (cold attenuation (CA) -15) and 30(cold attenuation (CA) -30) times by 10-day-old of commercial broiler chicks embryonated eggs at $29^{\circ}C$, respectively, The Physical and chemical properties (sensitivity to lipid solvents, low pH and thermostability), pathogenicity (mean death time, intracerebral pathogenic index and intravenous patho-genic index), safety, booster or protective effect and characterization of temperature sensitivity were measured in cold attenuated CA-15 or 30 strain and compared to those of parent CBP-1 strain. NDV, CBP-1 CA-30 strain acquired cold attenuation and decreased infectivity at $41^{\circ}C$ compared to those of parent strain grown at $37^{\circ}C$. It lost hemagglutination activity (HA) and cell infectivity at $56^{\circ}C$ for 30, 60, and 120 Min. CA-30 strain treated with ethyl ether also lost its HA and cell infectivity. Both CA-30 and parent strains exhibited a little resistant to HA at pH 3.0 glycine HCI buffer. Intracerebral pathogenic index (ICPI) and intravenous pathogenic index (IVPI) of parent strain were 1.12 and 1.45, but decreased to 0.75 and 0.00 in CA-30 strain, respectively. The safety was evaluated by mortality in chicks inoculated with 10$^{4.0}$ EID$_{50}$ /0.1 ml. The mortalities of parent, CA-30 and commercial Bl strains were 17.5, 12.0 and 0.0%, respectively. The safety of CA-30 strain was higher than that of parent strain. The booster effects of CA-30 strain and parent strain performed in 4-week-old chicks after being vaccinated with primary commercial Bl strain.