The FS-RNN incorporates the strengths of both multiscale RNNs and deep transition RNNs as it processes sequential data on different timescales and learns complex transition functions from one time step to the next.

Adaptive parameterization is a means of increasing this flexibility and thereby increasing the model's capacity to learn non-linear patterns.

This article presents a curated resource and evaluation suite for punctuation-aware treebank binarization. Standard binarization pipelines drop punctuation before head selection, which alters constituent shape and harms head-child identification. We release (1) a reproducible pipeline that preserves punctuation as sibling nodes prior to binarization, (2) derived artifacts and metadata (intermediate @X markers, reversibility signatures, alignment indices), and (3) an accompanying evaluation suite covering head-child prediction, round-trip reversibility, and structural compatibility with derivational resources (CCGbank). On the Penn Treebank, punctuation-aware preprocessing improves head prediction accuracy from 73.66\% (Collins rules) and 86.66\% (MLP) to 91.85\% with the same classifier, and achieves competitive alignment against CCGbank derivations. All code, configuration files, and documentation are released to enable replication and extension to other corpora.

In this study, we first tested the performance of the TreeTagger English model developed by Helmut Schmid with test files at our disposal, using this model to analyze relative clauses and noun complement clauses in English. We distinguished between the two uses of "that," both as a relative pronoun and as a complementizer. To achieve this, we employed an algorithm to reannotate a corpus that had originally been parsed using the Universal Dependency framework with the EWT Treebank. In the next phase, we proposed an improved model by retraining TreeTagger and compared the newly trained model with Schmid's baseline model. This process allowed us to fine-tune the model's performance to more accurately capture the subtle distinctions in the use of "that" as a complementizer and as a nominal. We also examined the impact of varying the training dataset size on TreeTagger's accuracy and assessed the representativeness of the EWT Treebank files for the structures under investigation. Additionally, we analyzed some of the linguistic and structural factors influencing the ability to effectively learn this distinction.

The development of lexicalized grammars, particularly Tree-Adjoining Grammar (TAG), has significantly advanced our understanding of syntax and semantics in natural language processing (NLP). While existing syntactic resources like the Penn Treebank and Universal Dependencies offer extensive annotations for phrase-structure and dependency parsing, there is a lack of large-scale corpora grounded in lexicalized grammar formalisms. To address this gap, we introduce TAGbank, a corpus of TAG derivations automatically extracted from existing syntactic treebanks. This paper outlines a methodology for mapping phrase-structure annotations to TAG derivations, leveraging the generative power of TAG to support parsing, grammar induction, and semantic analysis. Our approach builds on the work of CCGbank, extending it to incorporate the unique structural properties of TAG, including its transparent derivation trees and its ability to capture long-distance dependencies. We also discuss the challenges involved in the extraction process, including ensuring consistency across treebank schemes and dealing with language-specific syntactic idiosyncrasies. Finally, we propose the future extension of TAGbank to include multilingual corpora, focusing on the Penn Korean and Penn Chinese Treebanks, to explore the cross-linguistic application of TAG's formalism. By providing a robust, derivation-based resource, TAGbank aims to support a wide range of computational tasks and contribute to the theoretical understanding of TAG's generative capacity.

Processing sequential data of variable length is a major challenge in a wide range of applications, such as speech recognition, language modeling, generative image modeling and machine translation. Here, we address this challenge by proposing a novel recurrent neural network (RNN) architecture, the Fast-Slow RNN (FS-RNN). The FS-RNN incorporates the strengths of both multiscale RNNs and deep transition RNNs as it processes sequential data on different timescales and learns complex transition functions from one time step to the next. We evaluate the FS-RNN on two character level language modeling data sets, Penn Treebank and Hutter Prize Wikipedia, where we improve state of the art results to 1.19 and 1.25 bits-per-character (BPC), respectively. In addition, an ensemble of two FS-RNNs achieves 1.20 BPC on Hutter Prize Wikipedia outperforming the best known compression algorithm with respect to the BPC measure. We also present an empirical investigation of the learning and network dynamics of the FS-RNN, which explains the improved performance compared to other RNN architectures. Our approach is general as any kind of RNN cell is a possible building block for the FS-RNN architecture, and thus can be flexibly applied to different tasks.

Recurrent neural networks have flourished in many areas. Consequently, we can see new RNN cells being developed continuously, usually by creating or using gates in a new, original way. But what if we told you that gates in RNNs are redundant? In this paper, we propose a new recurrent cell called Residual Recurrent Unit (RRU) which beats traditional cells and does not employ a single gate. It is based on the residual shortcut connection, linear transformations, ReLU, and normalization. To evaluate our cell's effectiveness, we compare its performance against the widely-used GRU and LSTM cells and the recently proposed Mogrifier LSTM on several tasks including, polyphonic music modeling, language modeling, and sentiment analysis. Our experiments show that RRU outperforms the traditional gated units on most of these tasks. Also, it has better robustness to parameter selection, allowing immediate application in new tasks without much tuning.

In distributed training of deep neural networks, people usually run Stochastic Gradient Descent (SGD) or its variants on each machine and communicate with other machines periodically. However, SGD might converge slowly in training some deep neural networks (e.g., RNN, LSTM) because of the exploding gradient issue. Gradient clipping is usually employed to address this issue in the single machine setting, but exploring this technique in the distributed setting is still in its infancy: it remains mysterious whether the gradient clipping scheme can take advantage of multiple machines to enjoy parallel speedup. The main technical difficulty lies in dealing with nonconvex loss function, non-Lipschitz continuous gradient, and skipping communication rounds simultaneously. In this paper, we explore a relaxed-smoothness assumption of the loss landscape which LSTM was shown to satisfy in previous works, and design a communication-efficient gradient clipping algorithm. This algorithm can be run on multiple machines, where each machine employs a gradient clipping scheme and communicate with other machines after multiple steps of gradient-based updates. Our algorithm is proved to have $O\left(\frac{1}{N\epsilon^4}\right)$ iteration complexity and $O(\frac{1}{\epsilon^3})$ communication complexity for finding an $\epsilon$-stationary point in the homogeneous data setting, where $N$ is the number of machines. This indicates that our algorithm enjoys linear speedup and reduced communication rounds. Our proof relies on novel analysis techniques of estimating truncated random variables, which we believe are of independent interest. Our experiments on several benchmark datasets and various scenarios demonstrate that our algorithm indeed exhibits fast convergence speed in practice and thus validates our theory.