We present an automatic approach to tree annotation in which basic nonterminal symbols are alternately split and merged to maximize the likelihood of a training treebank. Starting with a simple Xbar grammar, we learn a new grammar whose nonterminals are subsymbols of the original nonterminals. In contrast with previous work, we are able to split various terminals to different degrees, as appropriate to the actual complexity in the data. Our grammars automatically learn the kinds of linguistic distinctions exhibited in previous work on manual tree annotation. On the other hand, our grammars are much more compact and substantially more accurate than previous work on automatic annotation. Despite its simplicity, our best grammar achieves an F1 of 90.2% on the Penn Treebank, higher than fully lexicalized systems.
We describe an algorithm for learning in the presence of multiple criteria. Our technique generalizes previous approaches in that it can learn optimal policies for all linear preference assignments over the multiple reward criteria at once. The algorithm can be viewed as an extension to standard reinforcement learning for MDPs where instead of repeatedly backing up maximal expected rewards, we back up the set of expected rewards that are maximal for some set of linear preferences (given by a weight vector, − → w ). We present the algorithm along with a proof of correctness showing that our solution gives the optimal policy for any linear preference function. The solution reduces to the standard value iteration algorithm for a specific weight vector, − → w .
To perform automatic, unconscious inference, the human brain must solve the "binding problem" by correctly grouping properties with objects. Temporal binding models like SHRUTI already suggest much of how this might be done in a connectionist and localist way by using temporal synchrony. We propose a set of alternatives to temporal synchrony mechanisms that instead use short signatures. This serves two functions: it allows us explore an additional biologically plausible alternative, and it allows us to extend and improve the capabilities of these models. These extensions model the human ability to both perform unification and handle multiple instantiations of logical terms. To verify our model's feasibility, we simulate it with a computer system modeling simple, neuron-like computations.
Figure 1: Defining a nonlinear projection of a 3D scene. Left: Original scene from a default view showing sketched 3D features. Right: The new, nonlinear projection. Two curve constraints are used to bow the side walls, and a point constraint is used to warp the back wall. A bstractLinear perspective is a good approximation to the format in which the human visual system conveys 3D scene information to the brain. Artists expressing 3D scenes, however, create nonlinear projections that balance their linear perspective view of a scene with elements of aesthetic style, layout and relative importance of scene objects. Manipulating the many parameters of a linear perspective camera to achieve a desired view is not easy. Controlling and combining multiple such cameras to specify a nonlinear projection is an even more cumbersome task. This paper presents a direct interface, where an artist manipulates in 2D the desired projection of a few features of the 3D scene. The features represent a rich set of constraints which define the overall projection of the 3D scene. Desirable properties of local linear perspective and global scene coherence drive a heuristic algorithm that attempts to interactively satisfy the given constraints as a weight-averaged projection of a minimal set of linear perspective cameras. This paper shows that 2D feature constraints are a direct and effective approach to control both the 2D layout of scene objects and the conceptually complex, high dimensional parameter space of nonlinear scene projection.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.