Text Processing Like Humans Do: Visually Attacking and Shielding NLP Systems
- Steffen Eger, Gözde Gül Sahin, Iryna Gurevych
- Computer ScienceNorth American Chapter of the Association for…
- 25 February 2019
This work investigates the impact of visual adversarial attacks on current NLP systems on character-, word-, and sentence-level tasks, showing that both neural and non-neural models are, in contrast to humans, extremely sensitive to such attacks, suffering performance decreases of up to 82%.
Dialogue Coherence Assessment Without Explicit Dialogue Act Labels
- Mohsen Mesgar, Sebastian Bucker, Iryna Gurevych
- Computer ScienceAnnual Meeting of the Association for…
- 22 August 2019
This work uses dialogue act prediction as an auxiliary task in a multi-task learning scenario to obtain informative utterance representations for coherence assessment, and alleviates the need for explicit dialogue act labels during evaluation.
Generating Coherent Summaries of Scientific Articles Using Coherence Patterns
- Daraksha Parveen, Mohsen Mesgar, M. Strube
- Computer ScienceConference on Empirical Methods in Natural…
- 1 November 2016
This work introduces a graph-based approach to summarize scientific articles using coherence patterns in a corpus of abstracts and proposes a method to combine coherence, importance and non-redundancy to generate the summary.
A Neural Local Coherence Model for Text Quality Assessment
- Mohsen Mesgar, M. Strube
- Computer ScienceConference on Empirical Methods in Natural…
- 17 August 2018
We propose a local coherence model that captures the flow of what semantically connects adjacent sentences in a text. We represent the semantics of a sentence by a vector and capture its state at…
Graph-based Coherence Modeling For Assessing Readability
- Mohsen Mesgar, M. Strube
- Computer ScienceInternational Workshop on Semantic Evaluation
- 1 June 2015
Novel graph-based coherence features based on frequent subgraphs are introduced and they outperform Pitler and Nenkova (2008) in the readability ranking task by more than 5% accuracy thus establishing a new state-of-the-art on this dataset.
Lexical Coherence Graph Modeling Using Word Embeddings
- Mohsen Mesgar, M. Strube
- Computer ScienceNorth American Chapter of the Association for…
- 1 June 2016
The lexical coherence graph (LCG), a new graph-based model to represent lexical relations among sentences, is introduced and Kneser-Ney smoothing is adapted to smooth subgraphs’ frequencies, which improves performance.
Reward Learning for Efficient Reinforcement Learning in Extractive Document Summarisation
- Yang Gao, Yang Gao, Christian M. Meyer, Mohsen Mesgar, Iryna Gurevych
- Computer ScienceInternational Joint Conference on Artificial…
- 27 May 2019
RELIS is proposed, a novel RL paradigm that learns a reward function with Learning-to-Rank (L2R) algorithms at training time and uses this reward function to train an input-specific RL policy at test time and it is proved that RELIS guarantees to generate near-optimal summaries with appropriate L2R and RL algorithms.
Improving Factual Consistency Between a Response and Persona Facts
- Mohsen Mesgar, Edwin Simpson, Iryna Gurevych
- Computer Science, PsychologyConference of the European Chapter of the…
- 2021
This work proposes to fine-tune neural models for response generation by reinforcement learning and an efficient reward function that explicitly captures the consistency between a response and persona facts as well as semantic plausibility.
A Neural Graph-based Local Coherence Model
- Mohsen Mesgar, Leonardo F. R. Ribeiro, Iryna Gurevych
- Computer ScienceConference on Empirical Methods in Natural…
- 2021
This work defines a new, efficient, and effective baseline for local coherence modeling, and evaluates the neural graph-based model for two benchmark coherence evaluation tasks: sentence ordering and summary coherence rating.
Generating Persona-Consistent Dialogue Responses Using Deep Reinforcement Learning
- Mohsen Mesgar, Edwin Simpson, Yue Wang, Iryna Gurevych
- Computer Science, PsychologyArXiv
- 30 April 2020
This work proposes a novel approach to train transformer-based dialogue agents using actor-critic reinforcement learning, and defines a new reward function to assess generated responses in terms of persona consistency, topic consistency, and fluency.
...
...