arXiv subjects

Jiajie Zou

Publications and source records attributed to Jiajie Zou.

5 recordsLinked to original sources

arXiv2023

Complex Mapping between Neural Response Frequency and Linguistic Units in Natural Speech

When listening to connected speech, human brain can extract multiple levels of linguistic units, such as syllables, words, and sentences. It has been hypothesized that the time scale of cortical activity encoding each linguistic unit is commensurate with the time scale of that linguistic unit in speech. Evidence for the hypothesis originally comes from studies using the frequency-tagging paradigm that presents each linguistic unit at a constant rate, and more recently extends to studies on natural speech. For natural speech, it is sometimes assumed that neural encoding of different levels of linguistic units is captured by the neural response tracking speech envelope in different frequency bands (e.g., around 1 Hz for phrases, around 2 Hz for words, and around 4 Hz for syllables). Here, we analyze the coherence between speech envelope and idealized responses, each of which tracks a single level of linguistic unit. Four units, i.e., phones, syllables, words, and sentences, are separately considered. It is shown that the idealized phone-, syllable-, and word-tracking responses all correlate with the speech envelope both around 3-6 Hz and below ~1 Hz. Further analyses reveal that the 1-Hz correlation mainly originates from the pauses in connected speech. The results here suggest that a simple frequency-domain decomposition of envelope-tracking activity cannot separate the neural responses to different linguistic units in natural speech.

q-bio.NC↗

arXiv2023

Human Attention during Goal-directed Reading Comprehension Relies on Task Optimization

The computational principles underlying attention allocation in complex goal-directed tasks remain elusive. Goal-directed reading, i.e., reading a passage to answer a question in mind, is a common real-world task that strongly engages attention. Here, we investigate what computational models can explain attention distribution in this complex task. We show that the reading time on each word is predicted by the attention weights in transformer-based deep neural networks (DNNs) optimized to perform the same reading task. Eye-tracking further reveals that readers separately attend to basic text features and question-relevant information during first-pass reading and rereading, respectively. Similarly, text features and question relevance separately modulate attention weights in shallow and deep DNN layers. Furthermore, when readers scan a passage without a question in mind, their reading time is predicted by DNNs optimized for a word prediction task. Therefore, attention during real-world reading can be interpreted as the consequence of task optimization.

cs.CL↗

arXiv2023

Acoustic correlates of the syllabic rhythm of speech: Modulation spectrum or local features of the temporal envelope

The syllable is a perceptually salient unit in speech. Since both the syllable and its acoustic correlate, i.e., the speech envelope, have a preferred range of rhythmicity between 4 and 8 Hz, it is hypothesized that theta-band neural oscillations play a major role in extracting syllables based on the envelope. A literature survey, however, reveals inconsistent evidence about the relationship between speech envelope and syllables, and the current study revisits this question by analyzing large speech corpora. It is shown that the center frequency of speech envelope, characterized by the modulation spectrum, reliably correlates with the rate of syllables only when the analysis is pooled over minutes of speech recordings. In contrast, in the time domain, a component of the speech envelope is reliably phase-locked to syllable onsets. Based on a speaker-independent model, the timing of syllable onsets explains about 24% variance of the speech envelope. These results indicate that local features in the speech envelope, instead of the modulation spectrum, are a more reliable acoustic correlate of syllables.

cs.SD↗

arXiv2022

PALRACE: Reading Comprehension Dataset with Human Data and Labeled Rationales

Pre-trained language models achieves high performance on machine reading comprehension (MRC) tasks but the results are hard to explain. An appealing approach to make models explainable is to provide rationales for its decision. To investigate whether human rationales can further improve current models and to facilitate supervised learning of human rationales, here we present PALRACE (Pruned And Labeled RACE), a new MRC dataset with human labeled rationales for 800 passages selected from the RACE dataset. We further classified the question to each passage into 6 types. Each passage was read by at least 26 human readers, who labeled their rationales to answer the question. It is demonstrated that models such as RoBERTa-large outperforms human readers in all 6 types of questions, including inference questions, but its performance can be further improved when having access to the human rationales. Simpler models and pre-trained models that are not fine-tuned based on the task benefit more from human rationales, and their performance can be boosted by more than 30% by rationales. With access to human rationales, a simple model based on the GloVe word embedding can reach the performance of BERT-base.

cs.CL↗

arXiv2021

Using Adversarial Attacks to Reveal the Statistical Bias in Machine Reading Comprehension Models

Pre-trained language models have achieved human-level performance on many Machine Reading Comprehension (MRC) tasks, but it remains unclear whether these models truly understand language or answer questions by exploiting statistical biases in datasets. Here, we demonstrate a simple yet effective method to attack MRC models and reveal the statistical biases in these models. We apply the method to the RACE dataset, for which the answer to each MRC question is selected from 4 options. It is found that several pre-trained language models, including BERT, ALBERT, and RoBERTa, show consistent preference to some options, even when these options are irrelevant to the question. When interfered by these irrelevant options, the performance of MRC models can be reduced from human-level performance to the chance-level performance. Human readers, however, are not clearly affected by these irrelevant options. Finally, we propose an augmented training method that can greatly reduce models' statistical biases.

cs.CL↗