626 citations · 974 across the 18 of their papers we have counts for
23 papers
MIDR: Enrichment-Augmented Indexing for Multimodal Document Retrieval
Debanjan Mahata, Atharva Tendle, Daniel Preotiuc-Pietro +2
Retrieval over visually rich documents has a representation problem: important content often lives in tables, charts, figures, and layout relations that plain OCR linearizes, corru…
Improving Instruct Models for Free: A Study on Partial Adaptation
Ozan İrsoy, Pengxiang Cheng, Jennifer L. Chen +3
Instruct models, obtained from various instruction tuning or post-training steps, are commonly deemed superior and more usable than their base counterpart. While the model gains in…
M3DocRAG: Multi-modal Retrieval is What You Need for Multi-page Multi-document Understanding
Jaemin Cho, Debanjan Mahata, Ozan Irsoy +2
Document visual question answering (DocVQA) pipelines that answer questions from documents have broad applications. Existing methods focus on handling single-page documents with mu…
MixCE: Training Autoregressive Language Models by Mixing Forward and Reverse Cross-Entropies
Shiyue Zhang, Shijie Wu, Ozan Irsoy +4
Autoregressive language models are trained by minimizing the cross-entropy of the model distribution Q relative to the data distribution P -- that is, minimizing the forward cross-…
BloombergGPT: A Large Language Model for Finance
Shijie Wu, Ozan Irsoy, Steven Lu +6
The use of NLP in the realm of financial technology is broad and complex, with applications ranging from sentiment analysis and named entity recognition to question answering. Larg…
Distillation of encoder-decoder transformers for sequence labelling
Marco Farina, Duccio Pappadopulo, Anant Gupta +3
Driven by encouraging results on a wide range of tasks, the field of NLP is experiencing an accelerated race to develop bigger language models. This race for bigger models has also…
Weakly Supervised Headline Dependency Parsing
Adrian Benton, Tianze Shi, Ozan İrsoy +1
English news headlines form a register with unique syntactic properties that have been documented in linguistics literature since the 1930s. However, headlines have received surpri…
PathFinder: Discovering Decision Pathways in Deep Neural Networks
Ozan İrsoy, Ethem Alpaydın
Explainability is becoming an increasingly important topic for deep neural networks. Though the operation in convolutional layers is easier to understand, processing becomes opaque…
Disentangling Online Chats with DAG-Structured LSTMs
Duccio Pappadopulo, Lisa Bauer, Marco Farina +2
Many modern messaging systems allow fast and synchronous textual communication among many users. The resulting sequence of messages hides a more complicated structure in which inde…
Diversity-Aware Batch Active Learning for Dependency Parsing
Tianze Shi, Adrian Benton, Igor Malioutov +1
While the predictive performance of modern statistical dependency parsers relies heavily on the availability of expensive expert-annotated treebank data, not all annotations contri…
Learning Syntax from Naturally-Occurring Bracketings
Tianze Shi, Ozan İrsoy, Igor Malioutov +1
Naturally-occurring bracketings, such as answer fragments to natural language questions and hyperlinks on webpages, can reflect human syntactic intuition regarding phrasal boundari…
Corrected CBOW Performs as well as Skip-gram
Ozan İrsoy, Adrian Benton, Karl Stratos
Mikolov et al. (2013a) observed that continuous bag-of-words (CBOW) word embeddings tend to underperform Skip-gram (SG) embeddings, and this finding has been reported in subsequent…
Semantic Role Labeling as Syntactic Dependency Parsing
Tianze Shi, Igor Malioutov, Ozan İrsoy
We reduce the task of (span-based) PropBank-style semantic role labeling (SRL) to syntactic dependency parsing. Our approach is motivated by our empirical analysis that shows three…
Dialogue Act Classification in Group Chats with DAG-LSTMs
Ozan İrsoy, Rakesh Gosangi, Haimin Zhang +6
Dialogue act (DA) classification has been studied for the past two decades and has several key applications such as workflow automation and conversation analytics. Researchers have…
On Expected Accuracy
Ozan İrsoy
We empirically investigate the (negative) expected accuracy as an alternative loss function to cross entropy (negative log likelihood) for classification tasks. Coupled with softma…
Dropout Regularization in Hierarchical Mixture of Experts
Ozan İrsoy, Ethem Alpaydın
Dropout is a very effective method in preventing overfitting and has become the go-to regularizer for multi-layer neural networks in recent years. Hierarchical mixture of experts i…
Continuously Constructive Deep Neural Networks
Ozan İrsoy, Ethem Alpaydın
Traditionally, deep learning algorithms update the network weights whereas the network architecture is chosen manually, using a process of trial and error. In this work, we propose…
Collective Entity Disambiguation with Structured Gradient Tree Boosting
Yi Yang, Ozan Irsoy, Kazi Shefaet Rahman
We present a gradient-tree-boosting-based structured learning model for jointly disambiguating named entities in a document. Gradient tree boosting is a widely used machine learnin…
Ask Me Anything: Dynamic Memory Networks for Natural Language Processing
Ankit Kumar, Ozan Irsoy, Peter Ondruska +6
Most tasks in natural language processing can be cast into question answering (QA) problems over language input. We introduce the dynamic memory network (DMN), a neural network arc…
Distributed Decision Trees
Ozan İrsoy, Ethem Alpaydın
Recently proposed budding tree is a decision tree algorithm in which every node is part internal node and part leaf. This allows representing every decision tree in a continuous pa…
Modeling Compositionality with Multiplicative Recurrent Neural Networks
Ozan İrsoy, Claire Cardie
We present the multiplicative recurrent neural network as a general model for compositional meaning in language, and evaluate it on the task of fine-grained sentiment analysis. We…
Autoencoder Trees
Ozan İrsoy, Ethem Alpaydın
We discuss an autoencoder model in which the encoding and decoding functions are implemented by decision trees. We use the soft decision tree where internal nodes realize soft mult…
Bidirectional Recursive Neural Networks for Token-Level Labeling with Structure
Ozan İrsoy, Claire Cardie
Recently, deep architectures, such as recurrent and recursive neural networks have been successfully applied to various natural language processing tasks. Inspired by bidirectional…