Korea, Republic of

This presentation is addressing the problem of prime-time television screenplay summarization based on the SummScreen Forever Dreaming dataset. Our approach utilizes extended Longformers combined with sketch supervision including categories specifically for scene descriptions. Our system was able to produce the shortest summaries out of all submissions. While some problems with factual consistency still remain, the system was scoring highest among competitors in the ROUGE and BERTScore evaluation categories.

COLING 2022

Long Input Dialogue Summarization with Sketch Supervision for Summarization of Primetime Television Transcripts

summscreen

summarization

dialogue

[![](https://assets.underline.io/uploads/markdown_image/1/image/81a3c0317d24f663b49b996875024d45.png)](https://aclanthology.org/volumes/2022.coling-1/)



### THE CONFERENCE WE KNOW AND WE WANT.
**COLING**, the International Conference on Computational Linguistics, is one of the premier conferences for the natural language processing and computational linguistics.

First established in 1965, the biennial COLING conference is held in diverse parts of the globe and attracts participants from both top-ranked research centers and emerging countries. Today, the most important developments in our field are taking place not only in universities and academic research institutes but also in industrial research departments including tech-startups. COLING provides opportunities for all these communities to showcase their exciting discovery.

In fall of 2022, COLING will be held in Gyeongju in a hybrid format. All participants can either present at the venue site or join virtually. As more people get vaccinated, we are happy to provide safer environments for our colleagues. We believe that COLING 2022 will be one of the conferences, free from the pandemic. The hybrid format gives presenters and sponsors a valuable opportunity to promote their companies in both an online and in-person venue. For the first time in a long time, customers can interact with their sponsor's products first-hand. The online venue, too, gives sponsors the chance to network with those unable to attend the in-person session.

The hybrid format ultimately lends itself to greater exposure for our sponsors: COLING2022 will let you reach more potential partners and customers than ever before!<br><br><br>

To gain access to this event page you are required to register.<br> Find more information and pay the registration fee on event the organizer’s website **[https://coling2022.org/reg](https://coling2022.org/reg)**

COLING, the International Conference on Computational Linguistics, is one of the premier conferences for the natural language processing and computational linguistics.

workshop paper

This presentation will discuss our shared task paper, which introduces a generalized AMR scheme to better model the long conversations in TV episodes. We hope to expose the diverse relations between concepts and complex dialogue structures so that they will be more accessible to transformer encoders. We will also discuss how the high-level information in a scene of a TV episode may guide the encoding of local tokens and utterances.

AMR-augmented Hierarchical Network for TV TranscriptSummarization

How to efficiently turn long documents into summaries with explorable structure? This has become an increasingly compelling question, given the rapid growth of document amount, length, and complexity. Existing summarization systems face new challenges for long documents: (1) Most systems only produce a generic paragraph-level summary, and cannot include all important points in a long document. (2) State-of-the-art summarizers rely on Transformer, whose time complexity is quadratic to the document length, and scales to hundreds of words at best.

In this talk, I first discuss the design of efficient attentions for long document summarization to handle more than 10k tokens while prior work only processes hundreds of words. I then present a new task: hierarchical question-summary generation, where we summarize the salient content in the source document into a hierarchy of questions and summaries, where each follow-up question inquires about the content of its parent question-summary pair. Learnable hierarchical biases are designed as part of the Transformer attention calculation, to adjust attention weights based on tokens’ relative positions with regard to the document structure. For both work, I will introduce new datasets to support experiments and discuss evaluation challenges.

Long Document Summarization using Efficient Attentions and Document Structure

Figurative language uses words in a way that deviates from their conventionally accepted definitions in order to convey a more complicated meaning or heightened effect. Figurative language generation can be very useful in many downstream tasks including fiction writing, essay writing, dialogue systems, etc. In this talk, I will first give an overview of figurative language generation and then introduce our recent work on a special case of figurative language generation, that is, hyperbole generation. We construct the first large-scale hyperbole corpus HYPO-XL and propose MOVER, an unsupervised approach to hyperbole generation.

Towards Figurative Language Generation

In this talk, we focus on video-to-text summarization and
investigate how to best utilize multimodal information for
summarizing long inputs (e.g., an hour-long TV show) into long
outputs (e.g.,~a multi-sentence summary). We extend
SummScreen (Chen et al., 2021), a dialogue summarization
dataset consisting of transcripts of TV episodes with reference
summaries, and create a multimodal variant by collecting
corresponding full-length videos. We incorporate multimodal
information into a pre-trained textual summarizer efficiently using
adapter modules augmented with a hierarchical structure while tuning
only~5\% of model parameters. Our experiments demonstrate that
multimodal information offers superior performance over more
memory-heavy and fully fine-tuned textual summarization methods.

Hierarchical3D Adapters for Long Video-to-text Summarization

This system description paper details TEAM UFAL’s approach for the SummScreen, TVMegasite subtask of the CreativeSumm shared task. The subtask deals with creat- ing summaries for dialogues from TV Soap operas. We utilized BART based pre-trained model fine-tuned on SamSum dialouge sum- marization dataset. Few examples from Au- toMin dataset and the dataset provided by the organizers were also inserted into the data as a few-shot learning objective. The additional data was manually broken into chunks based on different boundaries in summary and the di- alogue file. For inference we choose a similar strategy as the top-performing team at AutoMin 2021, where the data is split into chunks, either on [SCENE_CHANGE] or exceeding a pre- defined token length, to accommodate the max- imum token possible in the pre-trained model for one example. We implemented two differ- ent strategies as splits on [SCENE_CHANGE] did not necessarily mean having less than 1024 tokens in a segment.

BART and SamSum based few-shot approach for creative Summarization

A significant number of studies apply acoustic and linguistic characteristics of human speech as prominent markers of dementia and depression. However, studies on discriminating depression from dementia are rare. Co-morbid depression is frequent in dementia and these clinical conditions share many overlapping symptoms, but the ability to distinguish between depression and dementia is essential as depression is often curable. In this work, we investigate the ability of clustering approaches in distinguishing between depression and dementia from human speech. We introduce a novel aggregated dataset, which combines narrative speech data from multiple conditions, i.e., Alzheimer's disease, mild cognitive impairment, healthy control, and depression. We compare linear and non-linear clustering approaches and show that non-linear clustering techniques distinguish better between distinct disease clusters. Our interpretability analysis shows that the main differentiating symptoms between dementia and depression are acoustic abnormality, repetitiveness (or circularity) of speech, word finding difficulty, coherence impairment, and differences in lexical complexity and richness.

Data-driven Approach to Differentiating between Depression and Dementia from Noisy Speech and Language Data

Code-mixing (CM) is a frequently observed phenomenon on social media platforms in multilingual societies such as India. While the increase in code-mixed content on these platforms provides good amount of data for studying various aspects of code-mixing, the lack of automated text analysis tools makes such studies difficult. To overcome the same, tools such as language identifiers and parts of-speech (POS) taggers for analysing code-mixed data have been developed. One such tool is Named Entity Recognition (NER), an important Natural Language Processing (NLP) task, which is not only a subtask of Information Extraction, but is also needed for downstream NLP tasks such as semantic role labeling. While entity extraction from social media data is generally difficult due to its informal nature, code-mixed data further complicates the problem due to its informal, unstructured and incomplete information. 
In this work, we present the first ever corpus for Kannada-English code-mixed social media data with the corresponding named entity tags for NER. We provide strong baselines with machine learning classification models such as CRF, Bi-LSTM, and Bi-LSTM-CRF on our corpus with word, character, and lexical features.


"Kanglish alli names!" Named Entity Recognition for Kannada-English Code-Mixed Social Media Data

The tremendous increase in social media usage for sharing Television (TV) experiences has provided a unique opportunity in the Public Health and Marketing sectors to understand viewer engagement and attitudes through viewer-generated content on social media. However, this opportunity also comes with associated technical challenges. Specifically, given a televised event and related tweets about
this event, we need methods to effectively align these tweets and the corresponding event. 
In this paper, we consider the specific ecosystem of the Superbowl 2020 and map viewer tweets to advertisements they are referring to. Our proposed model, AdBERT, is an effective few-shot learning framework that can
handle the technical challenges of establishing ad-relatedness, class imbalance as well as the scarcity of labeled data. As part of this study, we have curated and developed two datasets that can prove useful for Social TV research: 1) dataset of ad-related tweets and 2) dataset of ad descriptions of Superbowl advertisements. Explaining connections to SentenceBERT, we describe the advantages of AdBERT that allow us to make the most out of a challenging and interesting dataset that we will open-source.

AdBERT: An Effective Few Shot Learning Framework for Aligning Tweets to Superbowl Advertisements

Gender analysis of Twitter can reveal important
socio-cultural differences between male and female
users. There has been a significant effort
to analyze and automatically infer gender in the
past for most widely spoken languages’ content,
however, to our knowledge very limited work
has been done for Arabic. In this paper, we
perform an extensive analysis of differences
between male and female users in the Arabic
Twitter-sphere. We study differences in user
engagement, topics of interest, and the gender
gap in professions. Along with gender analysis,
we also propose a method to infer gender
by utilizing usernames, profile pictures, tweets,
and networks of friends. In order to do so, we
manually annotated gender and locations for
∼167K Twitter accounts associated with ∼92K
user location, which we make publicly available.
1 Our proposed gender inference method
achieves an F1 score of 82.1% (47.3% higher
than the majority baseline). We also developed
a demo and made it publicly available.

https://alt.qcri.org/resources/ArabGend.zip
https://asad.qcri.org/

ArabGend: Gender Analysis and Inference on Arabic Twitter

Persona and Knowledge dual context open-domain chat is a novel dialogue generation task introduced recently. While Persona and Knowledge is each interesting context of open-domain dialogue, the combination of both has not been well studied. We tackle Persona-Knowledge identification and response generation tasks in this paper. We define Persona and Knowledge Dual Context Identification as the task to identify Persona and Knowledge jointly for a given dialogue. We develop novel multi-context retrieval method that utilizes all contexts of a dialogue simultaneously while also requiring limited training via zero-shot inference due to compatibility with neural Q & A models. Furthermore, we perform dialogue generation with various decoding techniques and illustrate crucial elements. We achieve SOTA across official metrics with 90%+ Grounding accuracy for both Persona and Knowledge Retrieval and 23.62 SacreBLEU score for Response Generation on Call For Customized Conversation benchmark.

Premium content

Long Input Dialogue Summarization with Sketch Supervision for Summarization of Primetime Television Transcripts

Downloads

Next from COLING 2022

AMR-augmented Hierarchical Network for TV TranscriptSummarization

Similar lecture

AMR-augmented Hierarchical Network for TV TranscriptSummarization