Papers by Julian Eisenschlos
MultiFiT: Efficient Multi-lingual Language Model Fine-tuning (D19-1)
Copied to clipboard
| Challenge: | Pretrained language models require unlabelled data for training, while cross-lingual models underperform on low-resource languages. |
| Approach: | They propose a multi-lingual language model fine-tuning to train and fine- tune language models efficiently in their own language. |
| Outcome: | The proposed method outperforms existing models on two widely used datasets on cross-lingual classification tasks. |
Selectively Answering Visual Questions (2024.findings-acl)
Copied to clipboard
| Challenge: | Large multi-modal models (LMMs) are capable of visual question answering (VQA) with unprecedented accuracy. |
| Approach: | They propose a calibration score that can be used to quantify uncertainty in visual question answering models. |
| Outcome: | The proposed calibration score is better calibrated than in text-only models for in-context learning. |
Open Domain Question Answering over Tables via Dense Retrieval (2021.naacl-main)
Copied to clipboard
| Challenge: | Recent advances in open-domain QA focus on retrieving textual passages . a retriever designed to handle tabular context can improve retrieval quality . |
| Approach: | They propose a tabular-based retrieval model that improves retrieval quality over a BERT-based retriever. |
| Outcome: | The proposed retriever improves retrieval quality with mined hard negatives over a BERT-based retriever. |
MatCha: Enhancing Visual Language Pretraining with Math Reasoning and Chart Derendering (2023.acl-long)
Copied to clipboard
Fangyu Liu, Francesco Piccinno, Syrine Krichene, Chenxi Pang, Kenton Lee, Mandar Joshi, Yasemin Altun, Nigel Collier, Julian Eisenschlos
| Challenge: | Visual language models that are pretraining on natural images or image-text pairs crawled from the web perform poorly on visual language tasks such as ChartQA and ChartQA. |
| Approach: | They propose to perform several pretraining tasks that cover plot deconstruction and numerical reasoning which are key capabilities in visual language modeling. |
| Outcome: | The proposed model outperforms state-of-the-art methods on benchmarks such as PlotQA and ChartQA by as much as 20%. |
Table-To-Text generation and pre-training with TabT5 (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models (LLMs) are limited when it comes to structured or semi-structured domains like tables. |
| Approach: | They propose an encoder-decoder model that generates natural language text based on tables and textual inputs. |
| Outcome: | TabT5 achieves 15% increase in sequence accuracy on spreadsheet formula prediction and data-to-text generation domains. |
Understanding tables with intermediate pre-training (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Textual entailment is well studied, but is less well studied for table enlargement . a new dataset of millions of examples is used to train the model . |
| Approach: | They adapt a table-based BERT model to recognize entailment from a dataset . they evaluate table pruning techniques as a pre-processing step to improve model efficiency . |
| Outcome: | The proposed model improves training and prediction efficiency at a moderate drop in accuracy. |
Leveraging Data Recasting to Enhance Tabular Reasoning (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Existing approaches to create tabular inference data are limited by human annotation and synthetic generation. |
| Approach: | They propose a framework for semi-automatically recasting tabular data to make use of both approaches. |
| Outcome: | The proposed framework can be used to build tabular NLI instances from five datasets. |
DePlot: One-shot visual language reasoning by plot-to-table translation (2023.findings-acl)
Copied to clipboard
Fangyu Liu, Julian Eisenschlos, Francesco Piccinno, Syrine Krichene, Chenxi Pang, Kenton Lee, Mandar Joshi, Wenhu Chen, Nigel Collier, Yasemin Altun
| Challenge: | Existing models for visual language reasoning require tens of thousands of training examples and their reasoning capabilities are limited. |
| Approach: | They propose a one-shot solution to visual language reasoning by combining plot-to-text translation and reasoning over the translated text into a modality conversion module. |
| Outcome: | The proposed method improves on human-written queries on plots and charts compared with a fine-tuned SOTA model on human queries. |
Do ever larger octopi still amplify reporting biases? Evidence from judgments of typical colour (2022.aacl-short)
Copied to clipboard
| Challenge: | Language models trained on text-only corpora have no direct access to the physical world and thus suffer from reporting bias. |
| Approach: | They investigate reporting bias from the perspective of colour in larger language models such as PaLM and GPT-3. |
| Outcome: | The proposed models outperform smaller models on the basis of colour and more closely track human judgements than smaller models. |
MATE: Multi-view Attention for Table Transformer Efficiency (2021.emnlp-main)
Copied to clipboard
| Challenge: | Tables are ubiquitous on the web, and are rich in information. |
| Approach: | They propose a sparse-attention Transformer architecture for modeling documents that contain large tables. |
| Outcome: | The proposed architecture scales linearly with respect to speed and memory, and can handle documents containing more than 8000 tokens with current accelerators. |
Fool Me Twice: Entailment from Wikipedia Gamification (2021.naacl-main)
Copied to clipboard
| Challenge: | a new dataset of entailment pairs is released to challenge the goal of examining arbitrary statements. |
| Approach: | They propose a multi-player game that challenges players to solve entailment pairs . the game is open source and encourages adversarial examples . |
| Outcome: | The proposed game lowers the number of examples that can be solved using "shortcuts" the game is open source and the code is available for free. |
Selectively Answering Ambiguous Questions (2023.emnlp-main)
Copied to clipboard
| Challenge: | Prior work has focused on the case in which the question is clear and the answer is unambiguous but possibly unknown. |
| Approach: | They propose to use a sampled set of questions to calibrate answers to ambiguous questions with varying model scales. |
| Outcome: | The results show that sampling-based confidence scores help calibrate answers to relatively unambiguous questions, with more dramatic improvements on ambiguous ones. |
Universal Self-Adaptive Prompting (2023.emnlp-main)
Copied to clipboard
| Challenge: | a hallmark of modern large language models is their impressive general zero-shot and few-shot abilities . however, zero- shot performances are weaker due to the lack of guidance and the difficulty of applying existing automatic prompt design methods in general tasks. |
| Approach: | They propose an automatic prompt design approach specifically tailored for zero-shot learning that categorizes a possible NLP task into one of three possible task types and then uses a selector to select the most suitable queries and zero- shot model-generated responses as pseudo-demonstrations. |
| Outcome: | The proposed approach is able to generalize ICL to zero-shot learning tasks while also allowing for a more efficient and efficient prompt design. |
MiQA: A Benchmark for Inference on Metaphorical Questions (2022.aacl-short)
Copied to clipboard
| Challenge: | a benchmark is proposed to assess the capability of large language models to reason with conventional metaphors. |
| Approach: | They propose to assess the capability of large language models to reason with conventional metaphors. |
| Outcome: | The proposed benchmark compares pre-trained models on binary-choice tasks with human models . the results show that human models perform better on the largest model, compared to small models based on the same task . |
DoT: An efficient Double Transformer for NLP tasks with tables (2021.findings-acl)
Copied to clipboard
| Challenge: | Recent studies have shown that transformer-based approaches to NLP tasks are slow and require computational and memory costs. |
| Approach: | They propose a transformer-based model that decomposes a problem into two sub-tasks and a pruning transformer that takes as input the pruning scores. |
| Outcome: | The proposed model improves training and inference time by at least 50% for a small drop in accuracy and also enables the model to maintain similar accuracy as slower baseline models. |
TaPas: Weakly Supervised Table Parsing via Pre-training (2020.acl-main)
Copied to clipboard
| Challenge: | Answering natural language questions over tables is often seen as a semantic parsing task. |
| Approach: | They propose an approach to question answering over tables without generating logical forms by selecting table cells and optionally applying a corresponding aggregation operator. |
| Outcome: | The proposed approach outperforms or rivals existing models on three different datasets and performs on par with the state-of-the-art on WikiSQL and WikiTQ. |
Faithful Chart Summarization with ChaTS-Pi (2024.acl-long)
Copied to clipboard
| Challenge: | Chart-to-summary generation can help explore data, communicate insights, and help the visually impaired. |
| Approach: | They propose a reference-free chart summarization metric for scoring faithfulness . they use human raters to fix and rank candidate summaries from any chart-summarization model . |
| Outcome: | The proposed metric scores the summarization faithfulness according to human ratings better than reference-based metrics, either learned or n-gram based. |