The red pieces are yours. They move like chess pawns (but not exactly): They can move one square forward to an empty square and they can capture diagonally. A player wins when one of their pieces advances to the other side of the board or the opponent is in a position with no possible moves. At the beginning of the game, you may also skip your move so that the computer moves first. There are no draws; One side will always win and the other will lose.
The computer starts by making random moves and learns as you play. Whenever it loses (or runs into a "dead-end" situation), it adds the previous position to a list of positions to avoid, and won't make moves that lead to any positions on that list. The more the computer plays, the better it gets. Set the AI mode to Pre-trained to play against a computer that already knows the best moves. Or set it to Tree Search to use an algorithm that looks ahead to find the best move.