1 code implementation • 22 Apr 2020 • Jurek Leonhardt, Avishek Anand, Megha Khosla
The extraction of main content from web pages is an important task for numerous applications, ranging from usability aspects, like reader views for news articles in web browsers, to information retrieval or natural language processing.
1 code implementation • 28 Jun 2022 • Mandeep Rathee, Thorben Funke, Avishek Anand, Megha Khosla
Given a GNN model, several interpretability approaches exist to explain GNN models with diverse (sometimes conflicting) evaluation methodologies.
1 code implementation • 18 May 2021 • Thorben Funke, Megha Khosla, Mandeep Rathee, Avishek Anand
In this paper, we lay down some of the fundamental principles that an explanation method for graph neural networks should follow and introduce a metric RDT-Fidelity as a measure of the explanation's effectiveness.
1 code implementation • 11 Jan 2021 • Zijian Zhang, Koustav Rudra, Avishek Anand
A desirable property of learning systems is to be both effective and interpretable.
1 code implementation • 12 Oct 2021 • Jurek Leonhardt, Koustav Rudra, Megha Khosla, Abhijit Anand, Avishek Anand
In this paper, we propose the Fast-Forward index -- a simple vector forward index that facilitates ranking documents using interpolation of lexical and semantic scores -- as a replacement for contextual re-rankers and dense indexes based on nearest neighbor search.
1 code implementation • 19 Oct 2020 • Jonas Wallat, Jaspreet Singh, Avishek Anand
We found that ranking models forget the least and retain more knowledge in their final layer.
1 code implementation • EMNLP (BlackboxNLP) 2020 • Jonas Wallat, Jaspreet Singh, Avishek Anand
We found that ranking models forget the least and retain more knowledge in their final layer compared to masked language modeling and question-answering.
1 code implementation • 14 Jun 2021 • Jurek Leonhardt, Fabian Beringer, Avishek Anand
Recently, pre-trained contextual models, such as BERT, have shown to perform well in language related tasks.
1 code implementation • 23 Jun 2021 • Jurek Leonhardt, Koustav Rudra, Avishek Anand
We introduce the Select-and-Rank paradigm for document ranking, where we first output an explanation as a selected subset of sentences in a document.
1 code implementation • 24 Mar 2024 • Maria Heuss, Maarten de Rijke, Avishek Anand
We evaluate RankingSHAP for commonly used learning-to-rank datasets to showcase the more nuanced use of an attribution method while highlighting the limitations of selection-based explanations.
1 code implementation • 7 Dec 2018 • Avishek Anand, Megha Khosla, Jaspreet Singh, Jan-Hendrik Zab, Zijian Zhang
In this paper, we propose a scalable approach to train word embeddings by partitioning the input space instead in order to scale to massive text corpora while not sacrificing the performance of the embeddings.
1 code implementation • 23 Jun 2022 • Yumeng Wang, Lijun Lyu, Avishek Anand
The aim of our algorithms is to add/replace a small number of tokens to a highly relevant or non-relevant document to cause a large rank demotion or promotion.
1 code implementation • 12 Jun 2023 • Jonas Wallat, Tianyi Zhang, Avishek Anand
To foster reproducibility, the code, as well as the data used in this paper, are openly available.
1 code implementation • 22 Jan 2024 • Jonas Wallat, Adam Jatowt, Avishek Anand
In this study, we aim to investigate the underlying limitations of general-purpose LLMs when deployed for tasks that require a temporal understanding.
no code implementations • EMNLP 2017 • Besnik Fetahu, Katja Markert, Avishek Anand
For a Wikipedia article, determining the \emph{citation span} of a citation, i. e. what content is covered by a citation, is important as it helps decide for which content citations are still missing.
no code implementations • 30 Mar 2017 • Besnik Fetahu, Katja Markert, Wolfgang Nejdl, Avishek Anand
An important editing policy in Wikipedia is to provide citations for added statements in Wikipedia pages, where statements can be arbitrary pieces of text, ranging from a sentence to a paragraph.
no code implementations • 30 Mar 2017 • Besnik Fetahu, Katja Markert, Avishek Anand
We propose a two-stage supervised approach for suggesting news articles to entity pages for a given state of Wikipedia.
no code implementations • 14 Jan 2017 • Tuan Tran, Claudia Niederée, Nattiya Kanhabua, Ujwal Gadiraju, Avishek Anand
In this work, we present a novel approach for timeline summarization of high-impact events, which uses entities instead of sentences for summarizing the event at each individual point in time.
no code implementations • 22 Oct 2018 • Megha Khosla, Jurek Leonhardt, Wolfgang Nejdl, Avishek Anand
We also unearth the limitations of evaluations on directed graphs in previous works and propose a clear strategy for evaluating link prediction and graph reconstruction in directed graphs.
no code implementations • 19 Mar 2019 • Megha Khosla, Vinay Setty, Avishek Anand
However, there is no common ground for systematic comparison of embeddings to understand their behavior for different graphs and tasks.
no code implementations • 20 Jun 2019 • Christian Otto, Matthias Springstein, Avishek Anand, Ralph Ewerth
Two modalities are often used to convey information in a complementary and beneficial manner, e. g., in online news, videos, educational resources, or scientific publications.
no code implementations • 15 Jul 2019 • Zeon Trevor Fernando, Jaspreet Singh, Avishek Anand
In image classification, the reference input tends to be a plain black image.
no code implementations • 11 Oct 2019 • Maximilian Idahl, Megha Khosla, Avishek Anand
In this paper we propose and study the novel problem of explaining node embeddings by finding embedded human interpretable subspaces in already trained unsupervised node representation embeddings.
no code implementations • 24 Apr 2020 • Rishiraj Saha Roy, Avishek Anand
The last few years have seen an explosion of research on the topic of automated question answering (QA), spanning the communities of information retrieval, natural language processing, and artificial intelligence.
no code implementations • 29 Apr 2020 • Jaspreet Singh, Zhenye Wang, Megha Khosla, Avishek Anand
In extensive quantitative experiments we show that our approach outperforms other model agnostic explanation approaches across pointwise, pairwise and listwise LTR models in validity while not compromising on completeness.
no code implementations • 18 May 2020 • Avishek Anand, Lawrence Cavedon, Matthias Hagen, Hideo Joho, Mark Sanderson, Benno Stein
Dagstuhl Seminar 19461 "Conversational Search" was held on 10-15 November 2019.
no code implementations • 1 Jan 2021 • Thorben Funke, Megha Khosla, Avishek Anand
Graph Neural Networks (GNNs) are a flexible and powerful family of models that build nodes' representations on irregular graph-structured data.
no code implementations • 18 Jan 2021 • Megha Khosla, Avishek Anand
We study the classical weighted perfect matchings problem for bipartite graphs or sometimes referred to as the assignment problem, i. e., given a weighted bipartite graph $G = (U\cup V, E)$ with weights $w : E \rightarrow \mathcal{R}$ we are interested to find the maximum matching in $G$ with the minimum/maximum weight.
Data Structures and Algorithms Discrete Mathematics Combinatorics
no code implementations • 18 Jan 2021 • Zijian Zhang, Jaspreet Singh, Ujwal Gadiraju, Avishek Anand
Are humans consistently better at selecting features that make image recognition more accurate?
1 code implementation • 30 Mar 2021 • Koustav Rudra, Zeon Trevor Fernando, Avishek Anand
However, the documents are longer than the passages and such document ranking models suffer from the token limitation (512) of BERT.
no code implementations • NAACL (TrustNLP) 2021 • Maximilian Idahl, Lijun Lyu, Ujwal Gadiraju, Avishek Anand
Post-hoc explanation methods are an important class of approaches that help understand the rationale underlying a trained model's decision.
no code implementations • 15 Jun 2021 • Michael Völske, Alexander Bondarenko, Maik Fröbe, Matthias Hagen, Benno Stein, Jaspreet Singh, Avishek Anand
We investigate whether one can explain the behavior of neural ranking models in terms of their congruence with well understood principles of document ranking by using established theories from axiomatic IR.
no code implementations • 23 Jun 2021 • Mandeep Rathee, Zijian Zhang, Thorben Funke, Megha Khosla, Avishek Anand
However, GNNs remain hard to interpret as the interplay between node features and graph structure is only implicitly learned.
no code implementations • 12 Sep 2021 • Zijian Zhang, Koustav Rudra, Avishek Anand
It is therefore important to conduct user studies to correct models' inference biases and improve the model in a life-long learning manner in the future according to the user feedback.
no code implementations • 3 May 2022 • Zijian Zhang, Vinay Setty, Avishek Anand
We introduce SparcAssist, a general-purpose risk assessment tool for the machine learning models trained for language tasks.
no code implementations • 7 Jul 2022 • Abhijit Anand, Jurek Leonhardt, Koustav Rudra, Avishek Anand
This paper proposes a simple yet effective method to improve ranking performance on smaller datasets using supervised contrastive learning for the document ranking problem.
no code implementations • 4 Nov 2022 • Avishek Anand, Lijun Lyu, Maximilian Idahl, Yumeng Wang, Jonas Wallat, Zijian Zhang
Explainable information retrieval is an emerging research area aiming to make transparent and trustworthy information retrieval systems.
no code implementations • 9 Nov 2022 • Jurek Leonhardt, Marcel Jahnke, Avishek Anand
Dual-encoder-based neural retrieval models achieve appreciable performance and complement traditional lexical retrievers well due to their semantic matching capabilities, which makes them a common choice for hybrid IR systems.
no code implementations • 28 Jun 2023 • Avishek Anand, Venktesh V, Abhijit Anand, Vinay Setty
Querying, conversing, and controlling search and information-seeking interfaces using natural language are fast becoming ubiquitous with the rise and adoption of large-language models (LLM).
no code implementations • 31 Aug 2023 • Abhijit Anand, Venktesh V, Vinay Setty, Avishek Anand
We find that there are two inherent limitations of using LLMs as query re-writers -- concept drift when using only queries as prompts and large inference costs during query processing.
no code implementations • 2 Oct 2023 • Simone Piaggesi, Megha Khosla, André Panisson, Avishek Anand
Towards that, we first develop new metrics that measure the global interpretability of embedding vectors based on the marginal contribution of the embedding dimensions to predicting graph structure.
no code implementations • 26 Oct 2023 • Venktesh V, Sourangshu Bhattacharya, Avishek Anand
We transfer the ability to decompose complex questions to simpler questions or generate step-by-step rationales to LLMs, by careful selection from available data sources of related tasks.
no code implementations • 2 Nov 2023 • Jurek Leonhardt, Henrik Müller, Koustav Rudra, Megha Khosla, Abhijit Anand, Avishek Anand
Dual-encoder-based dense retrieval models have become the standard in IR.
no code implementations • 26 Nov 2023 • Abhijit Anand, Jurek Leonhardt, Jaspreet Singh, Koustav Rudra, Avishek Anand
We then adapt a family of contrastive losses for the document ranking task that can exploit the augmented data to learn an effective ranking model.
no code implementations • 25 Mar 2024 • Venktesh V, Abhijit Anand, Avishek Anand, Vinay Setty
This addresses the challenge of verifying real-world numerical claims, which are complex and often lack precise information, not addressed by existing works that mainly focus on synthetic claims.
no code implementations • 3 Apr 2024 • Abhijit Anand, Venktesh V, Vinay Setty, Avishek Anand
In our extensive experiments on the DL-Hard dataset, we find that a principled query performance based scoring method using base and specialized ranker offers a significant improvement of up to 25% on the passage ranking task and up to 48. 4% on the document ranking task when compared to the baseline performance of using original queries, even outperforming SOTA model.