Shannon and Weaver categorized communications into three levels of problems: the technical problem, which tries to answer the question "how accurately can the symbols of communication be transmitted?"; the semantic problem, which asks the question "how precisely do the transmitted symbols convey the desired meaning?"; the effectiveness problem, which strives to answer the question "how effectively does the received meaning affect conduct in the desired way?". Traditionally, communication technologies mainly addressed the technical problem, ignoring the semantics or the effectiveness problems. Recently, there has been increasing interest to address the higher level semantic and effectiveness problems, with proposals ranging from semantic to goal oriented communications. In this thesis, we propose to formulate the semantic problem as a joint source-channel coding (JSCC) problem and the effectiveness problem as a multi-agent partially observable Markov decision process (MA-POMDP). As such, for the semantic problem, we propose DeepWiVe, the first-ever end-to-end JSCC video transmission scheme that leverages the power of deep neural networks (DNNs) to directly map video signals to channel symbols, combining video compression, channel coding, and modulation steps into a single neural transform. We also further show that it is possible to use predefined constellation designs as well as secure the physical layer communication against eavesdroppers for deep learning (DL) driven JSCC schemes, making such schemes much more viable for deployment in the real world. For the effectiveness problem, we propose a novel formulation by considering multiple agents communicating over a noisy channel in order to achieve better coordination and cooperation in a multi-agent reinforcement learning (MARL) framework. Specifically, we consider a MA-POMDP, in which the agents, in addition to interacting with the environment, can also communicate with each other over a noisy communication channel. The noisy communication channel is considered explicitly as part of the dynamics of the environment, and the message each agent sends is part of the action that the agent can take. As a result, the agents learn not only to collaborate with each other but also to communicate "effectively'' over a noisy channel. Moreover, we show that this framework generalizes both the semantic and technical problems. In both instances, we show that the resultant communication scheme is superior to one where the communication is considered separately from the underlying semantic or goal of the problem.