United Arab Emirates

Models able to generate free-text rationales that explain their output have been proposed as an important step towards interpretable NLP for “reasoning” tasks such as natural language inference and commonsense question answering. However, the relative merits of different architectures and types of rationales are not well understood and hard to measure. In this paper, we contribute two insights to this line of research: First, we find that models trained on gold explanations learn to rely on these but, in the case of the more challenging question answering data set we use, fail when given generated explanations at test time. However, additional fine-tuning on generated explanations teaches the model to distinguish between reliable and unreliable information in explanations. Second, we compare explanations by a generation-only model to those generated by a self-rationalizing model and find that, while the former score higher in terms of validity, factual correctness, and similarity to gold explanations, they are not more useful for downstream classification. We observe that the self-rationalizing model is prone to hallucination, which is punished by most metrics but may add useful context for the classification step.

EMNLP 2022

Human Ratings Do Not Reflect Downstream Utility: A Study of Free-Text Explanations for Model Predictions

free-text explanations

explanations

explainability

hallucinations

interpretability

gpt-2

bert

generation

workshop paper

## Welcome to EMNLP 2022!
I am delighted to welcome you to EMNLP 2022! I believe this conference has been complicated beyond any precedent. Over the past year, it’s been thrilling to see the organization team approach each new puzzle with creativity and enthusiasm. We hope that those participating in Abu Dhabi as well as those joining remotely will leave the conference feeling newly inspired by the program and newly connected to our ever-growing community. Following EMNLP 2021 and major NLP conferences since, EMNLP 2022 is “hybrid,” serving both virtual and in-person participants.

Our key innovations for EMNLP 2022 include:

* EMNLP 2022 is “hybrid” in a second sense, as well: we allowed both direct and rolling review paper submissions, building on the pilot experiment of EMNLP 2021, which considered a small number of ARR submissions. 
* Familiar from NAACL but new to EMNLP, we’ve added an industry track.
* During the conference, “portals” will link virtual poster sessions to in-person conference participants during poster sessions each day.
* The first ACL-family conference in the United Arab Emirates.

 *Message from Noah A. Smith, University of Washington and Allen Institute for AI, Seattle, Washington, USA* 
***EMNLP 2022 General Chair***
 
[![](https://assets.underline.io/uploads/markdown_image/1/image/9eec7d4a287ee18c278b08229290aa83.png)](https://drive.google.com/file/d/1OlPv6QBeo62VVTughj2jkiLeyHd1WnUt/view)
 
[![](https://assets.underline.io/uploads/markdown_image/1/image/a3db7a768409f05192210d98601edb25.png)](https://emnlp2022.rocket.chat/)

To access this site you need to register. Please register [here](https://2022.emnlp.org/registration/).

Register here

Welcome!
EMNLP 2022 will take place in Abu Dhabi from December 7th to December 11th, 2022. And it will be held in hybrid mode, both online and offline.

Acoustic word embeddings (AWEs) are fixed-dimensionality vector representations in a vector space such that different acoustic exemplars of the same word are projected nearby in the embedding space. In addition to their use in speech technology applications such as spoken term discovery and keyword spotting, AWE models have been adopted as models of spoken-word processing in several cognitively motivated studies and they have shown to exhibit a human-like performance in some auditory processing tasks. Nevertheless, the representation geometry of AWEs remains an under-explored topic that has not been studied in the literature. In this paper, we take a closer analytical look at AWEs and study how the choice of the learning objective and the architecture shapes their representational profile. Our main findings highlight the prominent role of the learning objective on the representational geometry over the architecture.

Analyzing the Representational Geometry of Acoustic Word Embeddings

We investigate how different domains are encoded in modern neural network architectures. We analyze the relationship between natural language domains, model size, and the amount of training data used. The primary analysis tool we develop is based on subpopulation analysis with Singular Vector Canonical Correlation Analysis (SVCCA), which we apply to Transformer-based language models (LMs). We compare the latent representations of such a language model at its different layers from a pair of models: a model trained on multiple domains (an experimental model) and a model trained on a single domain (a control model). Through our method, we find that increasing the model capacity impacts how domain information is stored in upper and lower layers differently. In addition, we show that larger experimental models simultaneously embed domain-specific information as if they were conjoined control models. These findings are confirmed qualitatively, demonstrating the validity of our method.

Understanding Domain Learning in Language Models Through Subpopulation Analysis

Interpretable entity representations (IERs) are sparse embeddings that are "human-readable'' in that dimensions correspond to fine-grained entity types and values are predicted probabilities that a given entity is of the corresponding type.  These methods perform well in zero-shot and low supervision settings.  Compared to standard dense neural embeddings, such interpretable representations may permit analysis and debugging.   However, while fine-tuning sparse, interpretable representations improves accuracy on downstream tasks, it destroys the semantics of the dimensions which were enforced in pre-training.  Can we maintain the interpretable semantics afforded by IERs while improving predictive performance on downstream tasks? Toward this end, we propose Intermediate enTity-based Sparse Interpretable Representation Learning (ItsIRL).   ItsIRL realizes improved performance over prior IERs on biomedical tasks, while maintaining "interpretability'' generally and their ability to support model debugging specifically.  The latter is enabled in part by the ability to perform "counterfactual'' fine-grained entity type manipulation, which we explore in this work.  Finally, we propose a method to construct entity type based class prototypes for revealing global semantic properties of classes learned by our model.  Code for pre-training and experiments will be made publicly available. 

Intermediate Entity-based Sparse Interpretable Representation Learning

Previous works on the fairness of toxic language classifiers compare the output of models with different identity terms as input features but do not consider the impact of other important concepts present in the context.  Here, besides identity terms, we take into account high-level latent features learned by the classifier and investigate the interaction between these features and identity terms. For a multi-class toxic language classifier, we leverage a concept-based explanation framework to calculate the sensitivity of the model to the concept of sentiment, which has been used before as a salient feature for toxic language detection. Our results show that although for some classes, the classifier has learned the sentiment information as expected, this information is outweighed by the influence of identity terms as input features. This work is a step towards evaluating procedural fairness, where unfair processes lead to unfair outcomes. The produced knowledge can guide debiasing techniques to ensure that important concepts besides identity terms are well-represented in training datasets.

Towards Procedural Fairness: Uncovering Biases in How a Toxic Language Classifier Uses Sentiment Information

Transformer based language models have been widely adopted by industrial and research organisations in developing machine learning applications in the presence of limited annotated data. While these models show remarkable results, their functioning in few-shot settings is still poorly understood. Hence, we perform an investigative study to understand the characteristics of such models fine-tuned in few-shot setups. Specifically, we compare the intermediate layer representations obtained from a few-shot model and a pre-trained language model. We observe that pre-trained and few-shot models show similar representations over initial layers, whereas the later layers show a stark deviation. Based on these observations, we propose to freeze the initial Transformer layers to fine-tune the model in a constrained text classification setup with $K$ annotated data points per class, where $K$ ranges from 8 to 64. In our experiments across six benchmark sentence classification tasks, we discover that freezing initial 50% Transformer layers not only reduces training time but also surprisingly improves Macro F1 (upto 8%) when compared to fully trainable layers in few-shot setup. We also observe that this idea of layer freezing can very well be generalized to state-of-the-art few-shot text classification techniques, like DNNC and LM-BFF, leading to significant reduction in training time while maintaining comparable performance.

Investigating the Characteristics of a Transformer in a Few-Shot Setup: Does Freezing Layers in RoBERTa Help?

We investigate to what extent a hundred publicly available, popular neural language models capture meaning systematically. Sentence embeddings obtained from pretrained or fine-tuned language models can be used to perform particular tasks, such as paraphrase detection, semantic textual similarity assessment or natural language inference. Common to all of these tasks is that paraphrastic sentences, that is, sentences that carry (nearly) the same meaning, should have (nearly) the same embeddings regardless of surface form. We demonstrate that performance varies greatly across different language models when a specific type of meaning-preserving transformation is applied: two sentences should be identified as paraphrastic if one of them contains a negated antonym in relation to the other one, such as "I am not guilty" versus "I am innocent". We introduce and release SemAntoNeg, a new test suite containing 3152 entries for probing paraphrasticity in sentences incorporating negation and antonyms. Among other things, we show that language models fine-tuned for natural language inference outperform other types of models, especially the ones fine-tuned to produce general-purpose sentence embeddings, on the test suite. Furthermore, we show that most models designed explicitly for paraphrasing are rather mediocre in our task.

It Is Not Easy To Detect Paraphrases: Analysing Semantic Similarity With Antonyms and Negation Using the New SemAntoNeg Benchmark

We propose a methodology and design two benchmark sets for measuring to what extent language-and-vision language models use the visual signal in the presence or absence of stereotypes. The first benchmark is designed to test for stereotypical colors of common objects, while the second benchmark considers gender stereotypes. The key idea is to compare predictions when the image conforms to the stereotype to predictions when it does not. Our results show that there is significant variation among multimodal models: the recent Transformer-based FLAVA seems to be more sensitive to the choice of image and less affected by stereotypes than older CNN-based models such as VisualBERT and LXMERT. This effect is more discernible in this type of controlled setting than in traditional evaluations where we do not know whether the model relied on the stereotype or the visual signal.

Controlling for Stereotypes in Multimodal Language Model Evaluation

Pretrained large generative language models have shown great performance on many tasks, but exhibit low compositional generalization abilities. Scaling such models has been shown to improve their performance on various NLP tasks even just by conditioning them on a few examples to solve the task without any fine-tuning (also known as in-context learning). In this work, we look at the gap between the in-distribution (ID) and out-of-distribution (OOD) performance of such models in semantic parsing tasks with in-context learning. In the ID settings, the demonstrations are from the same split ($\textit{test}$ or $\textit{train}$) that the model is being evaluated on, and in the OOD settings, they are from the other split. We look at how the relative generalization gap of in-context learning evolves as models are scaled up. We evaluate four model families, OPT, BLOOM, CodeGen and Codex on three semantic parsing datasets, CFQ, SCAN and GeoQuery with different number of exemplars, and observe a trend of decreasing relative generalization gap as models are scaled up.

On the Compositional Generalization Gap of In-Context Learning

Recent work has shown that neural feature- and representation-learning, e.g. BERT, achieves superior performance over traditional manual feature engineering based approaches, with e.g. SVMs, in translationese classification tasks. Previous research did not show $(i)$ whether the difference is because of the features, the classifiers or both, and $(ii)$ what the neural classifiers actually learn.  To address $(i)$, we carefully design experiments that swap features between BERT- and SVM-based classifiers. We show that an SVM fed with BERT representations performs at the level of the best BERT classifiers, while BERT learning and using handcrafted features performs at the level of an SVM using handcrafted features. This shows that the performance differences are due to the features. To address $(ii)$ we use integrated gradients and find that $(a)$ there is indication that information captured by hand-crafted features is only a subset of what BERT learns, and $(b)$ part of BERT's top performance results are due to BERT learning topic differences and spurious correlations with translationese.

Explaining Translationese: why are Neural Classifiers Better and what do they Learn?

GPT-3 has attracted much attention from both academia and industry. However, it is still unclear what GPT-3 has understood or learned especially in linguistic knowledge. Some studies have shown linguistic phenomena including negation and tense are hard to be recognized by language models such as BERT. In this study, we conduct probing tasks focusing on semantic information. Specifically, we investigate GPT-3's linguistic knowledge on semantic tasks to identify tense, the number of subjects, and the number of objects for a given sentence. We also experiment with different prompt designs and temperatures of the decoding method. Our experiment results suggest that GPT-3 has acquired linguistic knowledge to identify certain semantic information in most cases, but still fails when there are some types of disturbance happening in the sentence. We also perform error analysis to summarize some common types of mistakes that GPT-3 has made when dealing with certain semantic information.

Human Ratings Do Not Reflect Downstream Utility: A Study of Free-Text Explanations for Model Predictions

Downloads

Next from EMNLP 2022

Analyzing the Representational Geometry of Acoustic Word Embeddings

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES