arXiv subjects

Zhen Qin

Publications and source records attributed to Zhen Qin.

142 records · Page 8Linked to original sources

arXiv2022

ED2LM: Encoder-Decoder to Language Model for Faster Document Re-ranking Inference

State-of-the-art neural models typically encode document-query pairs using cross-attention for re-ranking. To this end, models generally utilize an encoder-only (like BERT) paradigm or an encoder-decoder (like T5) approach. These paradigms, however, are not without flaws, i.e., running the model on all query-document pairs at inference-time incurs a significant computational cost. This paper proposes a new training and inference paradigm for re-ranking. We propose to finetune a pretrained encoder-decoder model using in the form of document to query generation. Subsequently, we show that this encoder-decoder architecture can be decomposed into a decoder-only language model during inference. This results in significant inference time speedups since the decoder-only architecture only needs to learn to interpret static encoder embeddings during inference. Our experiments show that this new paradigm achieves results that are comparable to the more expensive cross-attention ranking approaches while being up to 6.8X faster. We believe this work paves the way for more efficient neural rankers that leverage large pretrained models.

cs.CL↗

arXiv2022

Improving Neural Ranking via Lossless Knowledge Distillation

We explore a novel perspective of knowledge distillation (KD) for learning to rank (LTR), and introduce Self-Distilled neural Rankers (SDR), where student rankers are parameterized identically to their teachers. Unlike the existing ranking distillation work which pursues a good trade-off between performance and efficiency, SDR is able to significantly improve ranking performance of students over the teacher rankers without increasing model capacity. The key success factors of SDR, which differs from common distillation techniques for classification are: (1) an appropriate teacher score transformation function, and (2) a novel listwise distillation framework. Both techniques are specifically designed for ranking problems and are rarely studied in the existing knowledge distillation literature. Building upon the state-of-the-art neural ranking structure, SDR is able to push the limits of neural ranking performance above a recent rigorous benchmark study and significantly outperforms traditionally strong gradient boosted decision tree based models on 7 out of 9 key metrics, the first time in the literature. In addition to the strong empirical results, we give theoretical explanations on why listwise distillation is effective for neural rankers, and provide ablation studies to verify the necessity of the key factors in the SDR framework.

cs.IR↗

arXiv2022

Charformer: Fast Character Transformers via Gradient-based Subword Tokenization

State-of-the-art models in natural language processing rely on separate rigid subword tokenization algorithms, which limit their generalization ability and adaptation to new settings. In this paper, we propose a new model inductive bias that learns a subword tokenization end-to-end as part of the model. To this end, we introduce a soft gradient-based subword tokenization module (GBST) that automatically learns latent subword representations from characters in a data-driven fashion. Concretely, GBST enumerates candidate subword blocks and learns to score them in a position-wise fashion using a block scoring network. We additionally introduce Charformer, a deep Transformer model that integrates GBST and operates on the byte level. Via extensive experiments on English GLUE, multilingual, and noisy text datasets, we show that Charformer outperforms a series of competitive byte-level baselines while generally performing on par and sometimes outperforming subword-based models. Additionally, Charformer is fast, improving the speed of both vanilla byte-level and subword-level Transformers by 28%-100% while maintaining competitive quality. We believe this work paves the way for highly performant token-free models that are trained completely end-to-end.

cs.CL↗

arXiv2022

cosFormer: Rethinking Softmax in Attention

Transformer has shown great successes in natural language processing, computer vision, and audio processing. As one of its core components, the softmax attention helps to capture long-range dependencies yet prohibits its scale-up due to the quadratic space and time complexity to the sequence length. Kernel methods are often adopted to reduce the complexity by approximating the softmax operator. Nevertheless, due to the approximation errors, their performances vary in different tasks/corpus and suffer crucial performance drops when compared with the vanilla softmax attention. In this paper, we propose a linear transformer called cosFormer that can achieve comparable or better accuracy to the vanilla transformer in both casual and cross attentions. cosFormer is based on two key properties of softmax attention: i). non-negativeness of the attention matrix; ii). a non-linear re-weighting scheme that can concentrate the distribution of the attention matrix. As its linear substitute, cosFormer fulfills these properties with a linear operator and a cosine-based distance re-weighting mechanism. Extensive experiments on language modeling and text understanding tasks demonstrate the effectiveness of our method. We further examine our method on long sequences and achieve state-of-the-art performance on the Long-Range Arena benchmark. The source code is available at https://github.com/OpenNLPLab/cosFormer.

cs.CL↗

arXiv2022

Are Pre-trained Convolutions Better than Pre-trained Transformers?

In the era of pre-trained language models, Transformers are the de facto choice of model architectures. While recent research has shown promise in entirely convolutional, or CNN, architectures, they have not been explored using the pre-train-fine-tune paradigm. In the context of language models, are convolutional models competitive to Transformers when pre-trained? This paper investigates this research question and presents several interesting findings. Across an extensive set of experiments on 8 datasets/tasks, we find that CNN-based pre-trained models are competitive and outperform their Transformer counterpart in certain scenarios, albeit with caveats. Overall, the findings outlined in this paper suggest that conflating pre-training and architectural advances is misguided and that both advances should be considered independently. We believe our research paves the way for a healthy amount of optimism in alternative architectures.

cs.CL↗

arXiv2021

Non-Clicks Mean Irrelevant? Propensity Ratio Scoring As a Correction

Recent advances in unbiased learning to rank (LTR) count on Inverse Propensity Scoring (IPS) to eliminate bias in implicit feedback. Though theoretically sound in correcting the bias introduced by treating clicked documents as relevant, IPS ignores the bias caused by (implicitly) treating non-clicked ones as irrelevant. In this work, we first rigorously prove that such use of click data leads to unnecessary pairwise comparisons between relevant documents, which prevent unbiased ranker optimization. Based on the proof, we derive a simple yet well justified new weighting scheme, called Propensity Ratio Scoring (PRS), which provides treatments on both clicks and non-clicks. Besides correcting the bias in clicks, PRS avoids relevant-relevant document comparisons in LTR training and enjoys a lower variability. Our extensive empirical evaluations confirm that PRS ensures a more effective use of click data and improved performance in both synthetic data from a set of LTR benchmarks, as well as in the real-world large-scale data from GMail search.

cs.IR↗

arXiv2021

OmniNet: Omnidirectional Representations from Transformers

This paper proposes Omnidirectional Representations from Transformers (OmniNet). In OmniNet, instead of maintaining a strictly horizontal receptive field, each token is allowed to attend to all tokens in the entire network. This process can also be interpreted as a form of extreme or intensive attention mechanism that has the receptive field of the entire width and depth of the network. To this end, the omnidirectional attention is learned via a meta-learner, which is essentially another self-attention based model. In order to mitigate the computationally expensive costs of full receptive field attention, we leverage efficient self-attention models such as kernel-based (Choromanski et al.), low-rank attention (Wang et al.) and/or Big Bird (Zaheer et al.) as the meta-learner. Extensive experiments are conducted on autoregressive language modeling (LM1B, C4), Machine Translation, Long Range Arena (LRA), and Image Recognition. The experiments show that OmniNet achieves considerable improvements across these tasks, including achieving state-of-the-art performance on LM1B, WMT'14 En-De/En-Fr, and Long Range Arena. Moreover, using omnidirectional representation in Vision Transformers leads to significant improvements on image recognition tasks on both few-shot learning and fine-tuning setups.

cs.CV↗

arXiv2021

Air-Ground Collaborative Mobile Edge Computing: Architecture, Challenges, and Opportunities

By pushing computation, cache, and network control to the edge, mobile edge computing (MEC) is expected to play a leading role in fifth generation (5G) and future sixth generation (6G). Nevertheless, facing ubiquitous fast-growing computational demands, it is impossible for a single MEC paradigm to effectively support high-quality intelligent services at end user equipments (UEs). To address this issue, we propose an air-ground collaborative MEC (AGC-MEC) architecture in this article. The proposed AGC-MEC integrates all potentially available MEC servers within air and ground in the envisioned 6G, by a variety of collaborative ways to provide computation services at their best for UEs. Firstly, we introduce the AGC-MEC architecture and elaborate three typical use cases. Then, we discuss four main challenges in the AGC-MEC as well as their potential solutions. Next, we conduct a case study of collaborative service placement for AGC-MEC to validate the effectiveness of the proposed collaborative service placement strategy. Finally, we highlight several potential research directions of the AGC-MEC.

cs.NI↗

arXiv2020

DeepKeyGen: A Deep Learning-based Stream Cipher Generator for Medical Image Encryption and Decryption

The need for medical image encryption is increasingly pronounced, for example to safeguard the privacy of the patients' medical imaging data. In this paper, a novel deep learning-based key generation network (DeepKeyGen) is proposed as a stream cipher generator to generate the private key, which can then be used for encrypting and decrypting of medical images. In DeepKeyGen, the generative adversarial network (GAN) is adopted as the learning network to generate the private key. Furthermore, the transformation domain (that represents the "style" of the private key to be generated) is designed to guide the learning network to realize the private key generation process. The goal of DeepKeyGen is to learn the mapping relationship of how to transfer the initial image to the private key. We evaluate DeepKeyGen using three datasets, namely: the Montgomery County chest X-ray dataset, the Ultrasonic Brachial Plexus dataset, and the BraTS18 dataset. The evaluation findings and security analysis show that the proposed key generation network can achieve a high-level security in generating the private key.

cs.CR↗

arXiv2020

Do RNN and LSTM have Long Memory?

The LSTM network was proposed to overcome the difficulty in learning long-term dependence, and has made significant advancements in applications. With its success and drawbacks in mind, this paper raises the question - do RNN and LSTM have long memory? We answer it partially by proving that RNN and LSTM do not have long memory from a statistical perspective. A new definition for long memory networks is further introduced, and it requires the model weights to decay at a polynomial rate. To verify our theory, we convert RNN and LSTM into long memory networks by making a minimal modification, and their superiority is illustrated in modeling long-term dependence of various datasets.

stat.ML↗

arXiv2019

An Improvement of PAA on Trend-Based Approximation for Time Series

Piecewise Aggregate Approximation (PAA) is a competitive basic dimension reduction method for high-dimensional time series mining. When deployed, however, the limitations are obvious that some important information will be missed, especially the trend. In this paper, we propose two new approaches for time series that utilize approximate trend feature information. Our first method is based on relative mean value of each segment to record the trend, which divide each segment into two parts and use the numerical average respectively to represent the trend. We proved that this method satisfies lower bound which guarantee no false dismissals. Our second method uses a binary string to record the trend which is also relative to mean in each segment. Our methods are applied on similarity measurement in classification and anomaly detection, the experimental results show the improvement of accuracy and effectiveness by extracting the trend feature suitably.

cs.LG↗

arXiv2018

Multi-Task Learning for Email Search Ranking with Auxiliary Query Clustering

User information needs vary significantly across different tasks, and therefore their queries will also differ considerably in their expressiveness and semantics. Many studies have been proposed to model such query diversity by obtaining query types and building query-dependent ranking models. These studies typically require either a labeled query dataset or clicks from multiple users aggregated over the same document. These techniques, however, are not applicable when manual query labeling is not viable, and aggregated clicks are unavailable due to the private nature of the document collection, e.g., in email search scenarios. In this paper, we study how to obtain query type in an unsupervised fashion and how to incorporate this information into query-dependent ranking models. We first develop a hierarchical clustering algorithm based on truncated SVD and varimax rotation to obtain coarse-to-fine query types. Then, we study three query-dependent ranking models, including two neural models that leverage query type information as additional features, and one novel multi-task neural model that views query type as the label for the auxiliary query cluster prediction task. This multi-task model is trained to simultaneously rank documents and predict query types. Our experiments on tens of millions of real-world email search queries demonstrate that the proposed multi-task model can significantly outperform the baseline neural ranking models, which either do not incorporate query type information or just simply feed query type as an additional feature.

cs.IR↗

arXiv2014

Document Clustering Based On Max-Correntropy Non-Negative Matrix Factorization

Nonnegative matrix factorization (NMF) has been successfully applied to many areas for classification and clustering. Commonly-used NMF algorithms mainly target on minimizing the $l_2$ distance or Kullback-Leibler (KL) divergence, which may not be suitable for nonlinear case. In this paper, we propose a new decomposition method by maximizing the correntropy between the original and the product of two low-rank matrices for document clustering. This method also allows us to learn the new basis vectors of the semantic feature space from the data. To our knowledge, we haven't seen any work has been done by maximizing correntropy in NMF to cluster high dimensional document data. Our experiment results show the supremacy of our proposed method over other variants of NMF algorithm on Reuters21578 and TDT2 databasets.

cs.IR↗

arXiv2013

Efficient Online Bootstrapping for Large Scale Learning

Bootstrapping is a useful technique for estimating the uncertainty of a predictor, for example, confidence intervals for prediction. It is typically used on small to moderate sized datasets, due to its high computation cost. This work describes a highly scalable online bootstrapping strategy, implemented inside Vowpal Wabbit, that is several times faster than traditional strategies. Our experiments indicate that, in addition to providing a black box-like method for estimating uncertainty, our implementation of online bootstrapping may also help to train models with better prediction performance due to model averaging.

cs.LG↗

arXiv2010

Treatment of Incompatible Initial and Boundary Data for Parabolic Equations in Higher Dimension

A new method is proposed to improve the numeri- cal simulation of time dependent problems when the initial and boundary data are not compatible. Unlike earlier methods limited to space dimension one, this method can be used for any space dimension. When both methods are applicable (in space dimen- sion one), the improvements in precision are comparable, but the method proposed here is not restricted by dimension.

math.NA↗

arXiv2010

Numerical Resolution near t = 0 of Nonlinear Evolution Equations in the Presence of Corner Singularities in Space Dimension 1

The incompatibilities between the initial and boundary data will cause singularities at the time-space corners, which in turn adversely affect the accuracy of the numerical schemes used to compute the solutions. We study the corner singularity issue for nonlinear evolution equations in 1D, and propose two remedy procedures that effectively recover much of the accuracy of the numerical scheme in use. Applications of the remedy procedures to the 1D viscous Burgers equation, and to the 1D nonlinear reaction-diffusion equation are presented. The remedy procedures are applicable to other nonlinear diffusion equations as well.

math.NA↗