TLDR: Extreme Summarization of Scientific Documents

By Isabel Cachola, Kyle Lo, Arman Cohan, Daniel S. Weld. 2020

291 citations recorded by the source corpus.

We introduce TLDR generation, a new form of extreme summarization, for scientific papers. TLDR generation involves high source compression and requires expert background knowledge and understanding of complex domain-specific language. To facilitate study on this task, we introduce SCITLDR, a new multi-target dataset of 5.4K TLDRs over 3.2K papers. SCITLDR contains both author-written and expert-derived TLDRs, where the latter are collected using a novel annotation protocol that produces high-quality summaries while minimizing annotation burden. We propose CATTS, a simple yet effective learning strategy for generating TLDRs that exploits titles as an auxiliary training signal. CATTS improves upon strong baselines under both automated metrics and human evaluations. Data and code are publicly available at https://github.com/allenai/scitldr.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Graham Neubig
2. Pengfei Liu
3. Weizhe Yuan
4. Iz Beltagy
5. Marek Rei
6. Lucy Lu Wang
7. Kris Cao
8. Jordan Clive

Citations in the corpus

• MSˆ2: Multi-Document Summarization of Medical Studies (2021, background): “…Li et al., 2016), relation extraction (Jain et al., 2020; Luan et al., 2018; Kringelum et al., 2016), QA (Abacha et al., 2019), NLI (Romanov and Shivade, 2018; Khot et al., 2018), summarization (Cachola et al., 2020; Chandrasekaran et al., 2019), claim verification (Wadden et al., 2020), and more.”
• Text Data Augmentation for Deep Learning (2021, unknown): “Cachola et al. [3] were able to collect a dataset of 5.4K Too Long; Didn’t Read (TLDR) summaries of 3.2K machine learning papers.”
• Control Prefixes for Text Generation (2021, methodology): “Cachola et al. (2020) demonstrate that control tokens can be used to effectively incorporate scaffold tasks alongside the main task for BARTLARGE .”
• Aspect-Controllable Opinion Summarization (2021, unknown): “Text summarization models, both extractive (Narayan et al., 2018; Zheng and Lapata, 2019; Cachola et al., 2020) and abstractive (See et al.”
• Control Prefixes for Parameter-Efficient Text Generation (2021, methodology): “This strategy was inspired by previous work which incorporates auxiliary scaffold tasks in multitask learning to improve span-labeling and text classification (Swayamdipta et al., 2018; Cohan et al., 2019; Cachola et al., 2020).”
• Can We Automate Scientific Reviewing? (2021, background): “…training samples compared with other benchmark summarization datasets, the few-shot learning ability of recent contextualized pre-trained models (Radford et al., 2019; Brown et al., 2020; Cachola et al., 2020a) still put training a passable review generation system from this dataset within grasp.”
• SummaC: Re-Visiting NLI-based Models for Inconsistency Detection in Summarization (2021, background): “Recent efforts to expand the application of summarization to new domains such as legal (Kornilova and Eidelman, 2019) or scholarly (Cachola et al., 2020) text will hopefully lead to the study of inconsistency detection in these novel domains, and perhaps even out of summarization on tasks such as te”
• Deep Literature Survey Automation with an Iterative Workflow (2025, unknown): “…refined rather than static (Jiang et al., 2023), the outline develops as an organically coherent structure rather than a patchwork (Zhang et al., 2025a), and paper cards enforce fine-grained evidence grounding that avoids distraction from peripheral details (Cachola et al., 2020; Wu et al., 2024).”