Ireland

Prompt-based probing has been widely used in evaluating the abilities of pretrained language models (PLMs). Unfortunately, recent studies have discovered such an evaluation may be inaccurate, inconsistent and unreliable. Furthermore, the lack of understanding its inner workings, combined with its wide applicability, has the potential to lead to unforeseen risks for evaluating and applying PLMs in real-world applications. To discover, understand and quantify the risks, this paper investigates the prompt-based probing from a causal view, highlights three critical biases which could induce biased results and conclusions, and proposes to conduct debiasing via causal intervention. This paper provides valuable insights for the design of unbiased datasets, better probing frameworks and more reliable evaluations of pretrained language models. Furthermore, our conclusions also echo that we need to rethink the criteria for identifying better pretrained language models.

ACL 2022

Can Prompt Probe Pretrained Language Models? Understanding the Invisible Risks from a Causal View

causal model

prompt-based probing

evaluation bias

pretrained language model

# Welcome everyone to ACL 2022!

The 60th Annual Meeting of the Association for Computational Linguistics is taking place May 22-27, 2022 as a hybrid event, in Dublin and online. We are happy to welcome all of you to this anniversary edition with an almost 50-50 in-person and virtual participation. 
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, three amazing keynote events and two new initiatives of invited talks: Spotlight Talks for Young Rising Stars (STIRS) and The Next Big Idea Talks. Posters (including Findings of ACL 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 will have an oral session and a poster session as part of Poster Session 1. The program also features eight Tutorials and 28 Workshops. 

 
We wish you a wonderful conference! 
[**The ACL 2022 Organizing Committee**](https://www.2022.aclweb.org/organisers)
 
[**Conference Handbook**](https://drive.google.com/file/d/1_BUCMfhMVrjG9E2e71aHdHeE28KSje0l/view?usp=sharing) 
[**Mini Handbook**](https://drive.google.com/file/d/1qlBKl0wzmlVF1oCeMQl3BahLd9nLP5Ce/view?usp=sharing) 
[**Posters and Demo guides**](https://drive.google.com/file/d/1UucMAoCNncIOaH1rMMDa0owuG9qgvJTG/view?usp=sharing)

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 (NLP). 

poster

Existing automatic evaluation systems of chatbots mostly rely on static chat scripts as ground truth, which is hard to obtain, and requires access to the models of the bots as a form of “white-box testing”. Interactive evaluation mitigates this problem but requires human involvement. During this presentation, we are going to present our work, an interactive chatbot evaluation framework in which chatbots compete with each other like in a sports tournament, using ﬂexible scoring metrics. This framework can efﬁciently rank chatbots independently from their model architectures and the domains for which they are trained.

ChatMatch: Evaluating Chatbots by Autonomous Chat Tournaments

Token-level adaptive training approaches can alleviate the token imbalance problem and thus improve neural machine translation, through re-weighting the losses of different target tokens based on specific statistical metrics (e.g., token frequency or mutual information). Given that standard translation models make predictions on the condition of previous target contexts, we argue that the above statistical metrics ignore target context information and may assign inappropriate weights to target tokens. While one possible solution is to directly take target contexts into these statistical metrics, the target-context-aware statistical computing is extremely expensive, and the corresponding storage overhead is unrealistic. To solve the above issues, we propose a target-context-aware metric, named conditional bilingual mutual information (CBMI), which makes it feasible to supplement target context information for statistical metrics. Particularly, our CBMI can be formalized as the log quotient of the translation model probability and language model probability by decomposing the conditional joint distribution. Thus CBMI can be efficiently calculated during model training without any pre-specific statistical calculations and large storage overhead. Furthermore, we propose an effective adaptive training approach based on both the token- and sentence-level CBMI. Experimental results on WMT14 English-German and WMT19 Chinese-English tasks show our approach can significantly outperform the Transformer baseline and other related methods.

Conditional Bilingual Mutual Information Based Adaptive Training for Neural Machine Translation

A desirable dialog system should be able to continually learn new skills without forgetting old ones, and thereby adapt to new domains or tasks in its life cycle. However, continually training a model often leads to a well-known catastrophic forgetting issue. In this paper, we present Continual Prompt Tuning, a parameter-efficient framework that not only avoids forgetting but also enables knowledge transfer between tasks. To avoid forgetting, we only learn and store a few prompt tokens' embeddings for each task while freezing the backbone pre-trained model. To achieve bi-directional knowledge transfer among tasks, we propose several techniques (continual prompt initialization, query fusion, and memory replay) to transfer knowledge from preceding tasks and a memory-guided technique to transfer knowledge from subsequent tasks. Extensive experiments demonstrate the effectiveness and efficiency of our proposed method on continual learning for dialog state tracking, compared with state-of-the-art baselines.

Continual Prompt Tuning for Dialog State Tracking

Example sentences for targeted words in a dictionary play an important role to help readers understand the usage of words. Traditionally, example sentences in a dictionary are usually created by linguistics experts, which are labor-intensive and knowledge-intensive. In this paper, we introduce the problem of dictionary example sentence generation, aiming to automatically generate dictionary example
sentences for targeted words according to the corresponding definitions. This task is challenging especially for polysemous words, because the generated sentences need to reflect different usages and meanings of these targeted words. Targeted readers may also have different backgrounds and educational levels. It is essential to generate example sentences that can be understandable for different backgrounds and levels of audiences. To solve these problems, we propose a controllable
target-word-aware model for this task. Our proposed model can generate reasonable examples for targeted words, even for polysemous words. In addition, our model allows users to provide explicit control over attributes related to readability, such as length and lexical complexity, thus generating suitable examples
for targeted audiences. Automatic and human evaluations on the Oxford dictionary dataset show that our model can generate suitable examples for targeted words with specific definitions while meeting the desired readability.

Controllable Dictionary Example Generation: Generating Example Sentences for Specific Targeted Audiences

Multilingual pre-trained language models, such as mBERT and XLM-R, have shown impressive cross-lingual ability. Surprisingly, both of them use multilingual masked language model (MLM) without any cross-lingual supervision or aligned data. Despite the encouraging results, we still lack a clear understanding of why cross-lingual ability could emerge from multilingual MLM. In our work, we argue that cross-language ability comes from the commonality between languages. Specifically, we study three language properties: constituent order, composition and word co-occurrence. First, we create an artificial language by modifying property in source language. Then we study the contribution of modified property through the change of cross-language transfer results on target language. We conduct experiments on six languages and two cross-lingual NLP tasks (textual entailment, sentence retrieval). Our main conclusion is that the contribution of constituent order and word co-occurrence is limited, while the composition is more crucial to the success of cross-linguistic transfer.

Cross-Lingual Ability of Multilingual Masked Language Models: A Study of Language Structure

Recently, contrastive learning has been shown to be effective in improving pre-trained language models (PLM) to derive high-quality sentence representations. It aims to pull close positive examples to enhance the alignment while push apart irrelevant negatives for the uniformity of the whole representation space. However, previous works mostly adopt in-batch negatives or sample from training data at random. Such a way may cause the sampling bias that improper negatives (e.g., false negatives and anisotropy representations) are used to learn sentence representations, which will hurt the uniformity of the representation space. To address it, we present a new framework DCLR (Debiased Contrastive Learning of unsupervised sentence Representations) to alleviate the influence of these improper negatives. In DCLR, we design an instance weighting method to punish false negatives and generate noise-based negatives to guarantee the uniformity of the representation space. Experiments on seven semantic textual similarity 
tasks show that our approach is more effective than competitive baselines. Our code and data are publicly available at the link: https://github.com/RUCAIBox/DCLR.

Debiased Contrastive Learning of Unsupervised Sentence Representations

Pre-training and fine-tuning is one of the main paradigms of current NLP tasks. Pre-trained models designed for specific tasks can achieve better performance than general pre-trained models, such as automatic summarization, dialogue response generation. There are already several pre-trained dialogue systems, for example, DialogGPT, Meena, Blenderbot and PLATO series. However, some challenges remain unresolved. There is a natural one-to-many problem in dialogue, that is, a single dialog context could be followed by multiple reasonable responses. Besides, existing pre-training architectures are often relatively simple, and the training objectives of pre-training and fine-tuning are inconsistent. On this basis, we propose DialogVED. In DialogVED, we introduces continuous latent variables. The model is pre-trained with large conversational datasets for multiple training objectives, and it is pre-trained in a fully supervised manner, which is closer to the fine-tuning stage.

DialogVED: A Pre-trained Latent Variable Encoder-Decoder Model for Dialog Response Generation

Although current state-of-the-art Transformer-based solutions succeeded in a wide range for single-document NLP tasks, they still struggle to address multi-input tasks such as multi-document summarization. Many solutions truncate the inputs, thus ignoring potential summary-relevant contents, which is unacceptable in the medical domain where each information can be vital. Others leverage linear model approximations to apply multi-input concatenation, worsening the results because all information is considered, even if it is conflicting or noisy with respect to a shared background. Despite the importance and social impact of medicine, there are no ad-hoc solutions for multi-document summarization. For this reason, we propose a novel discriminative marginalized probabilistic method (DAMEN) trained to discriminate critical information from a cluster of topic-related medical documents and generate a multi-document summary via token probability marginalization. Results prove we outperform the previous state-of-the-art on a biomedical dataset for multi-document summarization of systematic literature reviews. Moreover, we perform extensive ablation studies to motivate the design choices and prove the importance of each module of our method.

Discriminative Marginalized Probabilistic Neural Method for Multi-Document Summarization of Medical Literature

We consider event extraction in a generative manner with template-based conditional generation. Although there is a rising trend of casting the task of event extraction as a sequence generation problem with prompts, these generation-based methods have two significant challenges, including using suboptimal prompts and static event type information. In this paper, we propose a generative template-based event extraction method with dynamic prefix (GTEE-DynPref) by integrating context information with type-specific prefixes to learn a context-specific prefix for each context. Experimental results show that our model achieves competitive results with the state-of-the-art classification-based model OneIE on ACE 2005 and achieves the best performances on ERE. Additionally, our model is proven to be portable to new types of events effectively.

Dynamic Prefix-Tuning for Generative Template-based Event Extraction

While hyper-parameters (HPs) are important for knowledge graph (KG) learning, existing methods fail to search them efficiently. To solve this problem, we first analyze the properties of different HPs and measure the transfer ability from small subgraph to the full graph. Based on the analysis, we propose an efficient two-stage search algorithm KG- Tuner, which efficiently explores HP configurations on small subgraph at the first stage and transfers the top-performed configurations for fine-tuning on the large full graph at the second stage. Experiments show that our method can consistently find better HPs than the baseline algorithms within the same time budget, which achieves 9.1% average relative improvement for four embedding models on the large-scale KGs in open graph benchmark.

Can Prompt Probe Pretrained Language Models? Understanding the Invisible Risks from a Causal View

Downloads

Next from ACL 2022

ChatMatch: Evaluating Chatbots by Autonomous Chat Tournaments

Similar lecture

Position-based Prompting for Health Outcome Generation