Papers by Arpit Mittal
Simple Large-scale Relation Extraction from Unstructured Text (L18-1)
Copied to clipboard
| Challenge: | Knowledge-based question answering relies on the availability of facts, most of which cannot be found in structured sources. |
| Approach: | They propose a method for creating distant (weak) supervision labels for training a large-scale RE system by decoupling the model architecture from the feature design of a state-of-the-art neural network system. |
| Outcome: | The proposed method performs on par with the state-of-the-art model with similar features at 75x reduction in training time. |
Demand-Weighted Completeness Prediction for a Knowledge Base (N18-3)
Copied to clipboard
| Challenge: | Knowledge Bases (KBs) are widely used for representing information in a structured format. |
| Approach: | They propose a method to measure Demand-Weighted Completeness by defining an entity by its classes and using usage data to predict relation distributions. |
| Outcome: | The proposed method can be used to estimate completeness of knowledge bases based on how they are used and can quantify usage and completeness changes over time. |
Debiasing knowledge graph embeddings (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to train knowledge graph embeddings to be neutral to sensitive attributes such as gender have been shown to increase training time by a factor of eight or more. |
| Approach: | They propose a method where all embeddings are trained to be neutral to sensitive attributes such as gender by default using an adversarial loss. |
| Outcome: | The proposed method reduces training time by eightfold and improves accuracy. |
Large Scale Question Paraphrase Retrieval with Smoothed Deep Metric Learning (D19-55)
Copied to clipboard
| Challenge: | Question Paraphrase Retrieval (QPR) systems can be used to answer rare and noisy reformulations of common questions by mapping them to a set of canonical forms. |
| Approach: | They propose a Question Paraphrase Retrieval (QPR) system that retrieves equivalent questions that result in the same answer as the original question. |
| Outcome: | The proposed system outperforms the standard loss function in NIR with noisy labels on two QPR datasets. |
Learning When Not to Answer: a Ternary Reward Structure for Reinforcement Learning Based Question Answering (N19-2)
Copied to clipboard
| Challenge: | Existing methods for question answering over knowledge graphs use reinforcement learning to reason over a knowledge graph. |
| Approach: | They propose a new performance metric for question-answering agents that extends the binary reward structure to a ternary reward structure which rewards an agent for not answering a question rather than giving an incorrect answer. |
| Outcome: | The proposed method significantly improves the precision of answered questions while only not answering a limited number of correctly answered questions. |
FEVER: a Large-scale Dataset for Fact Extraction and VERification (N18-1)
Copied to clipboard
| Challenge: | 185,445 claims generated by altering sentences from Wikipedia are verified without knowledge of the sentence they were derived from. |
| Approach: | They propose a publicly available dataset for verification against textual sources, FEVER: Fact Extraction and VERification. |
| Outcome: | The proposed dataset achieves 31.87% accuracy on labeling a claim accompanied by the correct evidence, compared to 50.91% if we ignore the evidence. |
Generating Token-Level Explanations for Natural Language Inference (N19-1)
Copied to clipboard
| Challenge: | Existing methods to generate token-level explanations for NLI on single sentences have not been tested. |
| Approach: | They propose to generate token-level explanations for NLI without explicitly annotating training data. |
| Outcome: | The proposed approach is faster and more accurate than the black-box methods. |
Evaluating adversarial attacks against multiple fact verification systems (D19-1)
Copied to clipboard
| Challenge: | Automated fact verification is progressing due to advances in modeling and availability of large datasets. |
| Approach: | They propose two scoring metrics which take into account the correctness of adversarial instances. |
| Outcome: | The proposed method and paraphrasing method have higher potency and higher resilience than baselines. |
The FEVER2.0 Shared Task (D19-66)
Copied to clipboard
| Challenge: | Existing deep neural models are becoming more complex and difficult to understand and characterize their behaviour. |
| Approach: | They present the results of the second Fact Extraction and VERification (FEVER2.0) Shared Task. |
| Outcome: | The proposed task was based on the second Fact Extraction and VERification (FEVER2.0) shared task. |
Using Pairwise Occurrence Information to Improve Knowledge Graph Completion on Large-Scale Datasets (D19-1)
Copied to clipboard
| Challenge: | Existing methods for knowledge graph completion require large batch sizes and memory constraints. |
| Approach: | They combine occurrences of entity-relation pairs to construct a joint learning model using a dataset containing 2 million entities and combine them to increase the quality of sampled negatives. |
| Outcome: | The proposed model outperforms the baseline model on a dataset containing 2 million entities by 2.8% absolute on hits@1. |