United States

Intermediate layer knowledge distillation (KD) can improve the standard KD technique (which only targets the output of teacher and student models) especially over large pre-trained language models. However, intermediate layer distillation suffers from excessive computational burdens and engineering efforts required for setting up a proper layer  mapping. To address these problems, we propose a RAndom Intermediate Layer Knowledge Distillation (RAIL-KD) approach in which, intermediate layers from the teacher model are selected randomly to be distilled into the intermediate layers of the student model. This randomized selection enforces that all teacher layers are taken into account in the training process, while reducing the computational cost of intermediate layer distillation. Also, we show that it acts as a regularizer for improving the generalizability of the student model. We perform extensive experiments on GLUE tasks as well as on out-of-domain test sets. We show that our proposed RAIL-KD approach outperforms other state-of-the-art intermediate layer KD methods considerably in both performance and training-time.

NAACL 2022

RAIL-KD: RAndom Intermediate Layer Mapping for Knowledge Distillation

knowledge distillation

bert

EMNLP 2021 is planned to be a hybrid event in Punta Cana, Dominican Republic, with both on-site and fully virtual participation possible. The experience for on-site participants would closely approximate a normal pre-COVID *ACL conference, with 5-6 thematically organized parallel sessions and live Q/A and interactive discussion immediately after the talks. Presentations by virtual participants will be equitably interleaved with those of on-site participants, projected on the auditorium screens as if on-site, and also followed immediately by live Q/A and interactive discussion at a time during reasonable waking hours for the virtual presenter. For all participants, on-site and virtual, who are unable to attend a session due to either time-zone issues or because they are participating in another session live, talk recordings and slides will be available online at a minimum after the live presentation (and in many cases before as well), and questions may be submitted in advance on session-specific discussion boards and answered live in session with the usual visual aids if desired.

<iframe style="width:700px;height:400px" src="https://online.fliphtml5.com/ebtyf/ceby/" seamless="seamless" scrolling="no" frameborder="0" allowtransparency="true" allowfullscreen="true" ></iframe>

Please Note: The EMNLP registration system is not currently connected to the underline site as we are still in the process of building out EMNLP 2021. You will receive access instructions from underline the week of November 1st. 

Access is given only to EMNLP upon registration, if you have not registered please do so [here](https://2021.emnlp.org/registration).

Registered attendees will receive access the week of November 1st.

EMNLP 2021

EMNLP 2021 is planned to be a hybrid event in Punta Cana, Dominican Republic, with both on-site and fully virtual participation possible.

poster

## Welcome to NAACL 2022!
The Annual Meeting of the North American Chapter of the Association for Computational Linguistics (NAACL) is taking place July 10-15, 2022 as a hybrid event, in Seattle, WA and online. We are happy to welcome all of you to conference! 
 
The main conference program features oral presentations, in-person and virtual posters and demo sessions, a plenary session for our best paper presentations and awards, keynote presentations, a plenary panel on the place of linguistics and symbolic structures in NLP, and an Industrial Track panel on Careers in NLP.
 
Posters (including Findings of NAACL 2022) and demos are grouped by areas for both the in-person and the virtual sessions. For the virtual component, the talks will be on Zoom and the posters and the demos will be in GatherTown. The Student Research Workshop and will have an oral session and a poster session.
 
The program also features 6 Tutorials and 26 Workshops.
 
 We wish you a wonderful conference!
 
[The NAACL 2022 Organizing Committee](https://2022.naacl.org/committees/organization/)

2022 Annual Conference of the North American Chapter of the Association for Computational Linguistics

The Transformer architecture continues to show remarkable performance gains in many Natural Language Processing tasks. However, obtaining such state-of-the-art performance in different tasks requires fine-tuning the same model separately for each task. Clearly, such an approach is demanding in terms of both memory requirements and computing power. In this paper, aiming to improve training efficiency across multiple tasks, we propose to collectively factorize the weighs of the multi-head attention module of a pre-trained Transformer. We test our proposed method on finetuning multiple natural language understanding tasks by employing BERT-Large as an instantiation of the Transformer and the GLUE as the evaluation benchmark.  Experimental results show that our method requires training and storing only 1\% of the initial model parameters for each task and matches or improves the original fine-tuned model's performance for each task while effectively decreasing the parameter requirements by two orders of magnitude. Furthermore, compared to well-known adapter-based alternatives on the GLUE benchmark, our method consistently reaches the same levels of performance while requiring approximately four times fewer total and trainable parameters per task.

Efficient Learning of Multiple NLP Tasks via Collective Weight Factorization on BERT

Bias research in NLP is a rapidly growing and developing field.  
Similar to CrowS-Pairs (Nangia et al., 2020), we assess gender bias in masked-language models (MLMs) by studying pairs of sentences with gender swapped person references.
Most bias research focuses on and often is specific to English.
Using a novel methodology for creating sentence pairs that is applicable across languages, we create, based on CrowS-Pairs, a multilingual dataset for English, Finnish, German, Indonesian and Thai.
Additionally, we propose $S_JSD$, a new bias measure based on Jensen–Shannon divergence, which we argue retains more information from the model output probabilities than other previously proposed bias measures for MLMs.
Using multilingual MLMs, we find that $S_JSD$ diagnoses the same systematic biased behavior for non-English that previous studies have found for monolingual English pre-trained MLMs. 
$S_JSD$ outperforms the CrowS-Pairs measure, which struggles to find such biases for smaller non-English datasets.

An Information-Theoretic Approach and Dataset for Probing Gender Stereotypes in Multilingual Masked Language Models

Text style transfer is an important task in controllable language generation. Supervised approaches have pushed performance improvement on style-oriented rewriting such as formality conversion. However, challenges remain due to the scarcity of large-scale parallel data in many domains. While unsupervised approaches do not rely on annotated sentence pairs for each style, they are often plagued with instability issues such as mode collapse or quality degradation.
To take advantage of both supervised and unsupervised paradigms and tackle the challenges, in this work, we propose a semi-supervised framework for text style transfer. First, the learning process is bootstrapped with supervision guided by automatically constructed pseudo-parallel pairs using lexical and semantic-based methods. Then the model learns from unlabeled data via reinforcement rewards. Specifically, we propose to improve the sequence-to-sequence policy gradient via stepwise reward optimization, providing fine-grained learning signals and stabilizing the reinforced learning process. Experimental results show that the proposed approach achieves state-of-the-art performance on multiple datasets, and produces effective generation with as minimal as 10\% of training data.

Learning from Bootstrapping and Stepwise Reinforcement Reward: A Semi-Supervised Framework for Text Style Transfer

Syntax-controlled paraphrase generation aims to produce paraphrase conform to given syntactic patterns. To address this task, recent works have started to use parse trees (or syntactic templates) to guide generation.
A constituency parse tree contains abundant structural information, such as parent-child relation, sibling relation, and the alignment relation between words and nodes.
Previous works have only utilized parent-child and alignment relations, which may affect the generation quality.
To address this limitation, we propose a Structural Information-augmented Syntax-Controlled Paraphrasing (SI-SCP) model. Particularly, we design a syntax encoder based on tree-transformer to capture parent-child and sibling relations. To model the alignment relation between words and nodes, we propose an attention regularization objective, which makes the decoder accurately select corresponding syntax nodes to guide the generation of words.
Experiments show that SI-SCP achieves state-of-the-art performances in terms of semantic and syntactic quality on two popular benchmark datasets.
Additionally, we propose a Syntactic Template Retriever (STR) to retrieve compatible syntactic structures. 
We validate that STR is capable of retrieving compatible syntactic structures. We further demonstrate the effectiveness of SI-SCP to generate diverse paraphrases with retrieved syntactic structures.

Learning Structural Information for Syntax-Controlled Paraphrase Generation

Pre-trained models (PTMs) have lead to great improvements in natural language generation (NLG). However, it is still unclear how much commonsense knowledge they possess. With the goal of evaluating commonsense knowledge of NLG models, recent work has proposed the problem of generative commonsense reasoning, e.g., to compose a logical sentence given a set of unordered concepts. Existing approaches to this problem hypothesize that PTMs lack sufficient parametric knowledge for this task, which can be overcome by introducing external knowledge or task-specific pre-training objectives. Different from this trend, we argue that PTM's inherent ability for generative commonsense reasoning is underestimated due to the order-agnostic property of its input. In particular, we hypothesize that the order of the input concepts can affect the PTM's ability to utilize its commonsense knowledge. To this end, we propose a pre-ordering approach to elaborately manipulate the order of the given concepts before generation. Experiments show that our approach can outperform the more sophisticated models that have access to a lot of external data and resources. 

Revisiting Generative Commonsense Reasoning: A Pre-Ordering Approach

Existing multilingual video corpus moment retrieval (mVCMR) methods are mainly based on a two-stream structure. The visual stream utilizes the visual content in the video to estimate the query-visual similarity, and the subtitle stream exploits the query-subtitle similarity. The final query-video similarity ensembles similarities from two streams. In our work, we pro- pose a simple and effective strategy termed as Cross-lingual Cross-modal Consolidation (C3 ) to improve mVCMR accuracy. We adopt the ensemble similarity as the teacher to guide the training of each stream, leading to a more powerful ensemble similarity. Meanwhile, we use the teacher for a specific language to guide the student for another language to exploit the complementary knowledge across languages. Ex- tensive experiments on mTVR dataset demonstrate the effectiveness of our C3 method.

Cross-Lingual Cross-Modal Consolidation for Effective Multilingual Video Corpus Moment Retrieval

Self-supervised vision-and-language pretraining (VLP) aims to learn transferable multi-modal representations from large-scale image-text data and to achieve strong performances on a broad scope of vision-language tasks after finetuning. Previous mainstream VLP approaches typically adopt a two-step strategy relying on external object detectors to encode images in a multi-modal Transformer framework, which suffer from restrictive object concept space, limited image context and inefficient computation. In this paper,  we propose an object-aware end-to-end VLP framework, which directly feeds image grid features from CNNs into the Transformer and learns the multi-modal representations jointly. More importantly, we propose to perform object knowledge distillation to facilitate learning cross-modal alignment at different semantic levels. To achieve that, we design two novel pretext tasks by taking object features and their semantic labels from external detectors as supervision: 1.) Object-guided masked vision modeling task focuses on enforcing object-aware representation learning in the multi-modal Transformer; 2.) Phrase-region alignment task aims to improve cross-modal alignment by utilizing the similarities between noun phrases and object labels in the linguistic space. Extensive experiments on a wide range of vision-language tasks demonstrate the efficacy of our proposed framework, and we achieve competitive or superior performances over the existing pretraining strategies.

KD-VLP: Improving End-to-End Vision-and-Language Pretraining with Object Knowledge Distillation

Idioms are phrases which present a figurative meaning that cannot be (completely) derived by looking at the meaning of their individual components.
Identifying and understanding idioms in context is a crucial goal and a key challenge in a wide range of Natural Language Understanding tasks. Although efforts have been undertaken in this direction, the automatic identification and understanding of idioms is still a largely under-investigated area, especially when operating in a multilingual scenario.  
In this paper, we address such limitations and put forward several new contributions: we propose a novel multilingual Transformer-based system for the identification of idioms; we produce a high-quality automatically-created training dataset in 10 languages, along with a novel manually-curated evaluation benchmark; finally, we carry out a thorough performance analysis and release our evaluation suite at https://github.com/Babelscape/ID10M.

ID10M: Idiom Identification in 10 Languages

Named Entity Recognition (NER) is the task of identifying named entities in texts and classifying them through specific semantic categories, a process which is crucial for a wide range of NLP applications. Current datasets for NER focus mainly on coarse-grained entity types, tend to consider a single textual genre and to cover a narrow set of languages, thus limiting the general applicability of NER systems.
In this work, we design a new methodology for automatically producing NER annotations, and address the aforementioned limitations by introducing a novel dataset that covers 10 languages, 15 NER categories and 2 textual genres.
We also introduce a manually-annotated test set, and extensively evaluate the quality of our novel dataset on both this new test set and standard benchmarks for NER.
In addition, we include disambiguation information in our dataset in order to enable the development of multilingual entity linking systems.
We release our data and software at https://github.com/Babelscape/multinerd.

MultiNERD: A Multilingual, Multi-Genre and Fine-Grained Dataset for Named Entity Recognition (and Disambiguation)

Answering complex logical queries on incomplete knowledge graphs (KGs) with missing edges is a fundamental and important task for knowledge graph reasoning. The query embedding method is proposed to answer these queries by jointly encoding queries and entities to the same embedding space. Then the answer entities are selected according to the similarities between the entity embeddings and the query embedding. As the answers to a complex query are obtained from a combination of logical operations over sub-queries, the embeddings of the answer entities may not always follow a uni-modal distribution in the embedding space. Thus, it is challenging to simultaneously retrieve a set of diverse answers from the embedding space using a single and concentrated query representation such as a vector or a hyper-rectangle. To better cope with queries with diversified answers, we propose Query2Particles (Q2P), a complex KG query answering method. Q2P encodes each query into multiple vectors, named particle embeddings. By doing so, the candidate answers can be retrieved from different areas over the embedding space using the maximal similarities between the entity embeddings and any of the particle embeddings. Meanwhile, the corresponding neural logic operations are defined to support its reasoning over arbitrary first-order logic queries. The experiments show that Query2Particles achieves state-of-the-art performance on the complex query answering tasks on FB15k, FB15K-237, and NELL knowledge graphs. 

Downloads

Next from NAACL 2022

Efficient Learning of Multiple NLP Tasks via Collective Weight Factorization on BERT

Similar lecture

Universal-KD: Attention-based Output-Grounded Intermediate Layer Knowledge Distillation