Canada

The widespread adoption of large language models such as ChatGPT and Bard has led to unprecedented demand for these technologies. The burgeoning cost of inference for ever-increasing model sizes coupled with hardware shortages has limited affordable access and poses a pressing need for efficiency approaches geared towards high throughput and performance. Multi-input multi-output (MIMO) algorithms such as data multiplexing, offer a promising solution with a many-fold increase in throughput by performing inference for multiple inputs at the cost of a single input. Yet these approaches are not currently performant enough to be deployed in modern systems. We change that by developing MUX-PLMs, a class of high throughput pre-trained language models (PLMs) trained with data multiplexing, that can be fine-tuned for any downstream task to yield high-throughput high-performance. Our novel multiplexing and demultiplexing modules proficiently entangle and disentangle inputs, and enable high-performance high throughput \muxplms{} that are competitive with vanilla PLMs while achieving 2x/5x inference speedup with only a 1−4% drop on a broad suite of tasks.

ACL 2023

MUX-PLMs: Pre-training Language Models with Data Multiplexing

data multiplexing

efficient inference

pre-training

### Welcome to ACL 2023, the 61st Annual Meeting of the Association for Computational Linguistics! 
 The conference will be held in Toronto, Canada, July 9-14, 2023. 
Following the succession of the recent conferences in our field, ACL 2023 will adopt a hybrid format.
While the impact of Covid has considerably diminished in terms of traveling, obtaining visas to Canada
entails a very long process. Moreover, the global economic conditions pose challenges for many individuals to travel to conferences. Recognizing these circumstances, we know many participants may not be
able to attend the conference in person. Therefore, we are committed to providing a great virtual platform
so everyone has the opportunity to interact with other participants and enjoy the conference. Based on the
current registered participants, approxiately 30% have chosen to attend the conference virtually. Whether
you join us in person or virtually, we sincerely hope everyone has a remarkable conference experience. 
This General Chair’s message is where I express my gratitude to the many individuals who have made
enormous contributions to the conference over the past year.

Read [**ACL 2023 General Chair's message**](https://docs.google.com/document/d/1WobYM7norbG4dI48s75HfJoD89qgX5a_F-6U8AteLSA/edit?usp=sharing/) in full.

##### **[Conference Handbook](https://2023.aclweb.org/downloads/acl2023-handbook.pdf)**

The Association for Computational Linguistics (ACL) is the premier international scientific and professional society for people working on computational problems involving human language, a field often referred to as either computational linguistics or natural language processing.

workshop paper

Language models pretrained on general domain corpora usually exhibit considerable degradation when generalizing to downstream tasks of specialized domains. Existing approaches try to construct PLMs for each specific domains either from scratch or through further pretraining, which not only costs substantial resources, but also fails to cover all target domains at various granularity. In this work, we propose RADA, a novel Retrieval-Augmented framework for Domain Adaptation. We first construct a textual corpora that covers the downstream task at flexible domain granularity and resource availability. We employ it as a pluggable datastore to retrieve informative background knowledge, and integrate them into the standard language model framework to augment representations. We then propose a two-level selection scheme to integrate the most relevant information while alleviating irrelevant noises. Specifically, we introduce a differentiable sampling module as well as an attention mechanism to achieve both passage-level and word-level selection. Such a retrieval-augmented framework enables domain adaptation of language models with flexible domain coverage and fine-grained domain knowledge integration. We conduct comprehensive experiments across biomedical, science and legal domains to demonstrate the effectiveness of the overall framework, and its advantage over existing solutions.

Retrieval-Augmented Domain Adaptation of Language Models

Since the introduction of the SemEval 2020 Task 11, several approaches have been proposed in the literature for classifying propaganda
based on the rhetorical techniques used to influence readers.
These methods, however, classify one span at a time, ignoring dependencies from the labels of other spans within the same context.
In this paper, we approach propaganda technique classification as a
Multi-Instance Multi-Label (MIML) learning problem and propose a simple RoBERTa-based model for classifying all spans in an article simultaneously. Further, we note that, due to the annotation process where
annotators classified the spans by following a decision tree,
there is an inherent hierarchical relationship among the different
techniques, which existing approaches ignore. We incorporate these hierarchical label dependencies by adding an auxiliary classifier for each node in the decision tree to the training objective and ensembling the predictions from the original and auxiliary classifiers at test time. Overall, our model leads to an absolute improvement of 2.47% micro-F1 over the model from the shared task winning team in a cross-validation setup and is the best performing non-ensemble model on the shared task leaderboard.

Hierarchical Multi-Instance Multi-Label Learning for Detecting Propaganda Techniques

We investigate the problem of negation in
Pretrained Language Models. We decompose the
problem of negation into detecting / encoding
negation and actually using it in prediction and
probe for LM performance on them.
We also provide a theory on why PLMs perform
badly on negation.

Probing Negation in Language Models

Machine comprehension of procedural texts is essential for reasoning about the steps and automating the procedures. However, this requires identifying entities within a text and resolving the relationships between the entities. Previous work focused on the cooking domain and proposed a framework to convert a recipe text into a flow graph (FG) representation. In this work, we propose a framework based on the recipe FG for flow graph prediction of open-domain procedural texts. To investigate flow graph prediction performance in non-cooking domains, we introduce the wikiHow-FG corpus from articles on wikiHow, a website of how-to instruction articles. In experiments, we consider using the existing recipe corpus and performing domain adaptation from the cooking to the target domain. Experimental results show that the domain adaptation models achieve higher performance than those trained only on the cooking or target domain data.

Towards Flow Graph Prediction of Open-Domain Procedural Texts

The widespread usage of latent language representations via pre-trained language models (LMs) suggests that they are a promising source of structured knowledge. However, existing methods focus only on a single object per subject-relation pair, even though often multiple objects are correct. To overcome this limitation, we analyze these representations for their potential to yield materialized multi-object relational knowledge. We formulate the problem as a rank-then-select task. For ranking candidate objects, we evaluate existing prompting techniques and propose new ones incorporating domain knowledge. Among the selection methods, we find that choosing objects with a likelihood above a learned relation-specific threshold gives a 49.5% F1 score. Our results highlight the difficulty of employing LMs for the multi-valued slot-filling task and pave the way for further research on extracting relational knowledge from latent language representations.

Extracting Multi-valued Relations from Language Models

Recent models have shown that incorporating syntactic knowledge into the semantic role labelling (SRL) task leads to a significant improvement. In this paper, we propose Syntax-aware Graph-to-Graph Transformer (SynG2G-Tr) model, which encodes the syntactic structure using a novel way to input graph relations as embeddings, directly into the self-attention mechanism of Transformer. This approach adds a soft bias towards attention patterns that follow the syntactic structure but also allows the model to use this information to learn alternative patterns. We evaluate our model on both span-based and dependency-based SRL datasets, and outperform previous alternative methods in both in-domain and out-of-domain settings, on CoNLL 2005 and CoNLL 2009 datasets.

Syntax-Aware Graph-to-Graph Transformer for Semantic Role Labelling

Recent advances in prompt tuning have proven effective as a new language modeling paradigm for various natural language understanding tasks. However, it is challenging to adapt the soft prompt embeddings to different domains or generalize to low-data settings when learn- ing soft prompts itself is unstable, task-specific, and bias-prone. This paper proposes a prin- cipled learning framework—soft prompt con- struction (SPC)—to facilitate learning domain- adaptable soft prompts. Derived from the SPC framework is a simple loss that can plug into various models and tuning approaches to im- prove their cross-domain performance. We show SPC can improve upon SOTA for con- textual query rewriting, summarization, and paraphrase detection by up to 5%, 19%, and 16%, respectively.

SPC: Soft Prompt Construction for Cross Domain Generalization

When humans learn a new word, they can infer its grammatical properties from very few examples and apply this knowledge across grammatical contexts. Drawing inspiration from psycholinguistics, we conduct word-learning experiments to assess whether Long Short-Term Memory (LSTM) networks can achieve a similar level of grammatical generalisation. Although Transformer-based language models are successful few-shot learners, their complexity, size, and large data requirements render them less suitable for direct comparison with human language learning. Here, we test whether an LSTM, trained on a French corpus, can learn the grammatical gender of a previously unseen noun from a few examples in one agreement context and apply it to another. Our extensive experiments show that the LSTM effectively generalises grammatical knowledge. For example, the model learns the gender of a novel noun from a single example using gender-marked articles, 'le(m)' or 'la(f)', and applies this knowledge in a different context to predict noun-adjective and noun-participle gender agreement. This suggests that LSTMs do represent grammatical gender as a context-invariant property, similar to humans. Additionally, evaluating the biases and mechanisms that emerge from few-shot learning provides testable hypotheses for cognitive studies of human learning in relation to language models.

LSTMs Acquire More Than One Aspect of Gender with Few-shot Learning

Speech language pathologists rely on information spanning the layers of language, often drawing from multiple layers (e.g. phonology & semantics) at once. Recent innovations in large language models (LLMs) have been shown to build powerful representations for many complex language structures, especially syntax and semantics, unlocking the potential of large datasets through self-supervised learning techniques. However, these datasets are overwhelmingly orthographic, favoring writing systems like the English alphabet, a natural but phonetically imprecise choice. Meanwhile, LLM support for the international phonetic alphabet (IPA) ranges from poor to absent. Further, LLMs encode text at a word- or near-word level, and pre-training tasks have little to gain from phonetic/phonemic representations. In this paper, we introduce BORT, an LLM for mixed orthography/IPA meant to overcome these limitations. To this end, we extend the pre-training of an existing LLM with our own self-supervised pronunciation tasks. We then fine-tune for a clinical task that requires simultaneous phonological and semantic analysis. For an “easy” and “hard” version of these tasks, we show that fine-tuning from our models is more accurate by a relative 24% and 29%, and improved on character error rates by a relative 75% and 31%, respectively, than those starting from the original model.

Mixed Orthographic/Phonemic Language Modeling: Beyond Orthographically Restricted Transformers (BORT)

Although Question Answering (QA) have advanced to the human-level language skills in NLP tasks, there is still a problem: the QA model gets confused when there are similar sentences or paragraphs. Existing studies focus on enhancing the text understanding of the candidate answers to improve the overall performance of the QA models. However, since these methods focus on re-ranking queries or candidate answers, they fail to resolve the confusion when many generated answers are similar to the expected answer. To address these issues, we propose a novel contrastive learning framework called ContrastiveQA that alleviates the confusion problem in answer extraction. We propose a supervised method where we generate positive and negative samples from the candidate answers and the given answer, respectively. 
We thus introduce ContrastiveQA, which uses contrastive learning with sampling data to reduce incorrect answers. 
Experimental results on four QA benchmarks show the effectiveness of the proposed method.

Downloads

Next from ACL 2023

Retrieval-Augmented Domain Adaptation of Language Models

Similar lecture

PruMUX: Augmenting Data Multiplexing with Model Compression

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from ACL 2023

Retrieval-Augmented Domain Adaptation of Language Models

Similar lecture

PruMUX: Augmenting Data Multiplexing with Model Compression

Downloads