An Intelligent reflecting surface (IRS)-assisted wireless system is considered for simultaneous wireless information and power transfer (SWIPT) from a multi-antenna base station (BS) to single-antenna users. For this system, we formulate a multi-objective optimization problem (MOOP) to analyze the trade-off between the system energy efficiency (EE) and spectral efficiency (SE) by optimizing the beamformer at the BS, power splitting (PS) ratio at each user, and the RIS phase-shift taking into account the maximum transmit power, minimum harvested power, and the continuous phase-shift constraint. The investigated problem is a type of NP-hard problem that cannot be solved with traditional methods. Hence to solve the optimization problem, a learning-based solution based on Meta Deep Deterministic Policy Gradient (Meta-DDPG) method is presented to analyze the trade-off. Simulation results are presented to illustrate this trade-off under different number of users and weighting factor in the MOOP. Also, superiority of Meta-DDPG over DDPG, TD3, and Genetic Algorithm is illustrated.