Challenges of a small world analysis for the continuous monitoring of behavior in mice

Documenting a mouse's "real world" behavior in the "small world" of a laboratory cage with continuous video recordings offers insights into phenotypical expression of mouse genotypes, development and aging, and neurological disease. Nevertheless, there are challenges in the design of a small world, the behavior selected for analysis, and the form of the analysis used. Here we offer insights into small world analyses by describing how acute behavioral procedures can guide continuous behavioral methodology. We show how algorithms can identify behavioral acts including walking and rearing, circadian patterns of action including sleep duration and waking activity, and the organization of patterns of movement into home base activity and excursions, and how they are altered with aging. We additionally describe how specific tests can be incorporated within a mouse's living arrangement. We emphasize how machine learning can condense and organize continuous activity that extends over extended periods of time.

Learning to Predict Consequences as a Method of Knowledge Transfer in Reinforcement Learning

The reinforcement learning (RL) paradigm allows agents to solve tasks through trial-and-error learning. To be capable of efficient, long-term learning, RL agents should be able to apply knowledge gained in the past to new tasks they may encounter in the future. The ability to predict actions' consequences may facilitate such knowledge transfer. We consider here domains where an RL agent has access to two kinds of information: agent-centric information with constant semantics across tasks, and environment-centric information, which is necessary to solve the task, but with semantics that differ between tasks. For example, in robot navigation, environment-centric information may include the robot's geographic location, while agent-centric information may include sensor readings of various nearby obstacles. We propose that these situations provide an opportunity for a very natural style of knowledge transfer, in which the agent learns to predict actions' environmental consequences using agent-centric information. These predictions contain important information about the affordances and dangers present in a novel environment, and can effectively transfer knowledge from agent-centric to environment-centric learning systems. Using several example problems including spatial navigation and network routing, we show that our knowledge transfer approach can allow faster and lower cost learning than existing alternatives.