We propose to create a novel predoctoral training program at the interface of Computer Science and Cognitive Science at Cornell University. While this interface has been crowded lately with "fuzzy-engineering" approaches typically identified as Artificial Intelligence (common-sense reasoning, neural nets, etc.), we plan to focus, instead, on "formal-scientific" aspects, including the formal modeling of computational and inference processes, computer- human interface and computer science education studies, vision and robotics, learning theory, and intelligent information retrieval. We argue that 1. both disciplines, Computer Science and Cognitive Science, will benefit from such a program, and 2. Cornell is uniquely suited to accommodate such a program. Indeed, we hope that the proposed program would serve as a model of how and engineering field and a liberal-arts field can leverage each other and ultimately lead to both (1) engineering artifices that pay more attention to the human user and (2) a better formal understanding of human thought process The proposal is organized in four parts. First, we present several con=crete and promising research directions. Each module is concrete in that at least one current doctoral student or faculty member is interested in pursuing the topic; each module is promising in that progress will depend on and advance both participating disciplines. Second, we briefly portray the current infrastructure at Cornell and we present the structure of the proposed program. Third, we present a plan for recruiting and, fourth, we justify the need for funding.
