The need for high scalability of personalised instruction has been a constant demand in higher education, which is often not met by traditional lectures. The path to adaptive learning environments that customize instruction to the individual student lies with cognitive computing systems that sense, reason, learn, and adapt through machine learning, natural-language processing, and knowledge representation. The most developed realization of this vision is the intelligent tutoring system (ITS), which integrates a model of the subject, a model of the learner and a model of pedagogy to help students solve problems with individualized hints and feedback. This paper reviews the cognitive computing for adaptive learning with the intelligent tutoring in higher education background. It brings together the field along four-axis: the four parts of the ITS architecture, the adaptive loop that connects them; the cognitive-computing techniques that enable adaptivity, including Bayesian and deep knowledge tracing, content adaptation and sequencing, natural-language dialogue and automated feedback, and affect and engagement detection; the applications and evidence of effectiveness in mathematics, programming, science, language, and writing instruction; and the open challenges of scalability, data privacy, generalisability, equity, and evaluation. The literature reviewed indicates that there is consistently evidence, based on meta-analysis, that intelligent tutoring is close to the effectiveness of human one-to-one tutoring and significantly more effective than conventional instruction, provided implementation is robust and fidelity is high, and provided measures are appropriate. The single central thesis is that the utility of cognitive computing in education is not any one specific technique, rather it is the precision of the learner model upon which the teaching changes are based; and that the potential contribution of this utility in higher education will rely on integration, assessment, and equity as much as algorithmic progress.