Knowledge distillation (KD) is a popular method of transferring knowledge from a large "teacher" model to a small "student" model. KD can be divided into two categories: prediction matching and intermediate-layer matching. We explore an intriguing phenomenon: layer-selection strategy does not matter (much) in intermediate-layer matching. In this paper, we show that seemingly nonsensical matching strategies such as matching the teacher's layers in reverse still result in surprisingly good student performance. We provide an interpretation for this phenomenon by examining the angles between teacher layers viewed from the student's perspective.

Large language models have been increasing in size due to their success in a wide range of applications. This calls for a pressing need to reduce memory usage to make them more accessible. Post-training quantization is a popular technique which uses fewer bits (e.g., 4--8 bits) to represent the model without retraining it. However, it remains a challenging task to perform quantization in an ultra-low-bit setup (e.g., 2 bits). In this paper, we propose InvarExplore, a unified framework that systematically explores different model invariance at the same time, allowing us to take advantage of the synergy between each type of invariance. Importantly, InvarExplore features a discrete search algorithm that enables us to explore permutation invariance, which is under-studied as it cannot be optimized with gradient-based methods. Results show that InvarExplore is compatible with existing state-of-the-art methods, achieving an add-on performance improvement over strong competing methods.

Machine translation is a widely applicable NLP task that translates a text from a source language to a target language Brown et al. (1990); Bahdanau et al. (2015). The Transformer architecture Vaswani et al. (2017) and pretrained large language models Radford et al. (2019); Raffel et al. (2020); Lewis et al. (2020) have largely improved translation performance, especially in the supervised setting, where a model can learn from large volumes of parallel corpora. However, machine translation remains challenging for low-resource languages, because there are not enough data for large neural networks to learn these languages. We specifically focus on multilingual translation in the zero-shot setting, where the system is required to translate between unseen language pairs. Since collecting parallel data and training individual models for every translation pair are prohibitively expensive, it is common to build a single multilingual system Johnson et al. (2017); Fan et al. (2021) that can perform translation for all language pairs, most of which are zero-shot translation directions with few exceptions (e.g., English). These models work by prepending a language-indicator token, and zero-shot ability emerges as the model generalizes from trained language pairs to unseen ones (Liu et al., 2021; Wicks and Duh, 2022).

We address unsupervised discontinuous constituency parsing, where we observe a high variance in the performance of the only previous model. We propose to build an ensemble of different runs of the existing discontinuous parser by averaging the predicted trees, to stabilize and boost performance. To begin with, we provide comprehensive computational complexity analysis (in terms of P and NP-complete) for tree averaging under different setups of binarity and continuity. We then develop an efficient exact algorithm to tackle the task, which runs in a reasonable time for all samples in our experiments. Results on three datasets show our method outperforms all baselines in all metrics; we also provide in-depth analyses of our approach.

Knowledge distillation (KD) is the process of transferring knowledge from a large model to a small one. It has gained increasing attention in the natural language processing community, driven by the demands of compressing ever-growing language models. In this work, we propose an f-DISTILL framework, which formulates sequence-level knowledge distillation as minimizing a generalized f-divergence function. We propose four distilling variants under our framework and show that existing SeqKD and ENGINE approaches are approximations of our f-DISTILL methods. We further derive step-wise decomposition for our f-DISTILL, reducing intractable sequence-level divergence to word-level losses that can be computed in a tractable manner. Experiments across four datasets show that our methods outperform existing KD approaches, and that our symmetric distilling losses can better force the student to learn from the teacher distribution.

Open-domain dialogue systems aim to interact with humans through natural language texts in an open-ended fashion. Despite the recent success of super large dialogue systems such as ChatGPT, using medium-to-small-sized dialogue systems remains the common practice as they are more lightweight and accessible; however, generating diverse dialogue responses is challenging, especially with smaller models. In this work, we propose an Equal-size Hard Expectation--Maximization (EqHard-EM) algorithm to train a multi-decoder model for diverse dialogue generation. Our algorithm assigns a sample to a decoder in a hard manner and additionally imposes an equal-assignment constraint to ensure that all decoders are well-trained. We provide detailed theoretical analysis to justify our approach. Further, experiments on two large-scale open-domain dialogue datasets verify that our EqHard-EM algorithm generates high-quality diverse responses.

Open-domain dialogue systems aim to converse with humans through text, and its research has heavily relied on benchmark datasets. In this work, we first identify the overlapping problem in DailyDialog and OpenSubtitles, two popular open-domain dialogue benchmark datasets. Our systematic analysis then shows that such overlapping can be exploited to obtain fake state-of-the-art performance. Finally, we address this issue by cleaning these datasets and setting up a proper data processing procedure for future research.