Human one-to-one tutoring has been shown to be a very effective form of instruction. Three contrasting hypotheses, a tutor-centered one, a student-centered one, and an interactive one could all potentially explain the effectiveness of tutoring. To test these hypotheses, analyses focused not only on the effectiveness of the tutors' moves, but also on the effectiveness of the students' construction on learning, as well as their interaction. The interaction hypothesis is further tested in the second study by manipulating the kind of tutoring tactics tutors were permitted to use. In order to promote a more interactive style of dialogue, rather than a didactic style, tutors were suppressed from giving explanations and feedback. Instead, tutors were encouraged to prompt the students. Surprisingly, students learned just as effectively even when tutors were suppressed from giving explanations and feedback. Their learning in the interactive style of tutoring is attributed to construction from deeper and a greater amount of scaffolding episodes, as well as their greater effort to take control of their own learning by reading more. What they learned from reading was limited, however, by their reading abilities.

For many forms of e-learning environments, the system's behavior can be viewed as a sequential decision process wherein, at each discrete step, the system is responsible for selecting the next action to take. Pedagogical strategies are policies to decide the next system action when there are multiple ones available. In this project we present a Reinforcement Learning (RL) approach for inducing effective pedagogical strategies and empirical evaluations of the induced strategies. This paper addresses the technical challenges in applying RL to Cordillera, a Natural Language Tutoring System teaching students introductory college physics. The algorithm chosen for this project is a model-based RL approach, Policy Iteration, and the training corpus for the RL approach is an exploratory corpus, which was collected by letting the system make random decisions when interacting with real students. Overall, our results show that 123 138 M. Chi et al. by using a rather small training corpus, the RL-induced strategies indeed measurably improved the effectiveness of Cordillera in that the RL-induced policies improved students' learning gains significantly.

Type I restriction endonucleases are intriguing, multifunctional complexes that restrict DNA randomly, at sites distant from the target sequence. Restriction at distant sites is facilitated by ATP hydrolysis-dependent, translocation of double-stranded DNA towards the stationary enzyme bound at the recognition sequence. Following restriction, the enzymes are thought to remain associated with the DNA at the target site, hydrolyzing copious amounts of ATP. As a result, for the past 35 years type I restriction endonucleases could only be loosely classified as enzymes since they functioned stoichiometrically relative to DNA. To further understand enzyme mechanism, a detailed analysis of DNA cleavage by the EcoR124I holoenzyme was done. We demonstrate for the first time that type I restriction endonucleases are not stoichiometric but are instead catalytic with respect to DNA. Further, the mechanism involves formation of a dimer of holoenzymes, with each monomer bound to a target sequence and, following cleavage, each dissociates in an intact form to bind and restrict subsequent DNA molecules. Therefore, type I restriction endonucleases, like their type II counterparts, are true enzymes. The conclusion that type I restriction enzymes are catalytic relative to DNA has important implications for the in vivo function of these previously enigmatic enzymes.

Modeling patient disease progression using Electronic Health Records (EHRs) is critical to assist clinical decision making. Long-Short Term Memory (LSTM) is an effective model to handle sequential data, such as EHRs, but it encounters two major limitations when applied to EHRs: it is unable to interpret the prediction results and it ignores the irregular time intervals between consecutive events. To tackle these limitations, we propose an attention-based time-aware LSTM Networks (ATTAIN), to improve the interpretability of LSTM and to identify the critical previous events for current diagnosis by modeling the inherent time irregularity. We validate ATTAIN on modeling the progression of an extremely challenging disease, septic shock, by using real-world EHRs. Our results demonstrate that the proposed framework outperforms the state-of-the-art models such as RETAIN and T-LSTM. Also, the generated interpretative time-aware attention weights shed some lights on the progression behaviors of septic shock.

No abstract