"... In PAGE 5: ...imilarly, MIPS 16.19.01 in D2 is lower than GO 0017076 in DU, and so on. Therefore the integrated user data DU could look like in Table4 , where the semantic correspondences have been applied. In general, the user may want to answer queries such as the number of human proteins that are involved in kinase activity from the integrated data or even to infer models based on the data available in order to use them to predict useful information about new unlabeled data (e.... ..."

"... In PAGE 5: ...imilarly, MIPS 16.19.01 in D2 is lower than GO 0017076 in DU, and so on. Therefore the integrated user data DU could look like in Table4 , where the semantic correspondences have been applied. In general, the user may want to answer queries such as the number of human proteins that are involved in kinase activity from the integrated data or even to infer models based on the data available in order to use them to predict useful information about new unlabeled data (e.... ..."

"... In PAGE 4: ... Label Description initiate Begin new DU, content separate from previous uncompleted DUs continue same agentadds related content to open DU acknowledge Demonstrate or claim understanding of previous material by other agent repair Correct (potential) misunderstandingof DU content Request Repair Signal lack of understanding Request Ack Signal for other to acknowledge cancel Stop work on DU, leaving it un- grounded and ungroundable Table 2: Grounding Acts Acts are also distinguished as to whether they are per- formed by the initiator of a unit (I) or the responder (R)5 As well as the three states in the transition net- work in Figure 1, representing starting grounding (S), ac- ceptance/acknowledgment needed for grounding (1), and grounded material (F), the model in (Traum amp; Allen 1992), contains four other states, representing the need for a repair by initiator and acknowledgment by responder (2), need for acknowledgment by initiator (3), need for repair by re- sponder and acknowledgment by initiator (4), and a dead state (D), in which the unit was deemed to be ungroundable (though the same content could, of course be reintroduced and grounded in a subsequent unit). Table3 shows the tran- sition diagram, given each state and possible groundingacts (superscripts represent the performing agent, with respect to the initiator of this unit). A couple of other remarks on this model are in order.... In PAGE 5: ...g., the transitions from F to F, given an ack, in Table3 ). Likewise, until a fixed num- ber of subsequent DUs have been opened, it is also possible to repair a unit previously thought to be grounded, so the F state is a current best guess about what was grounded, rather than a final, irrevocable determination.... ..."

"... In PAGE 6: ... We segmented the enhancing tumor area, the region that appears hyper-intense after injecting the contrast agent (we also included non-enhancing areas contained within the enhancing contour). Table1 and 2 present Jaccard scores of testing DU i and DS i for each study, pi, respectively. While the standard supervised DRF improves over its degenerate model LR by 1%, semi-supervised DRF signi cantly improves over the supervised DRF by 11%, which is signi cant at p lt; 0:00566 using a paired example t test.... ..."