Pankaj Gupta

Soumen Basu, Mayank Gupta, Pratyaksha Rana et al.

We propose a novel deep neural network architecture to learn interpretable representation for medical image analysis. Our architecture generates a global attention for region of interest, and then learns bag of words style deep feature embeddings with local attention. The global, and local feature maps are combined using a contemporary transformer architecture for highly accurate Gallbladder Cancer (GBC) detection from Ultrasound (USG) images. Our experiments indicate that the detection accuracy of our model beats even human radiologists, and advocates its use as the second reader for GBC diagnosis. Bag of words embeddings allow our model to be probed for generating interpretable explanations for GBC detection consistent with the ones reported in medical literature. We show that the proposed model not only helps understand decisions of neural network models but also aids in discovery of new visual features relevant to the diagnosis of GBC. Source-code and model will be available at https://github.com/sbasu276/RadFormer

Soumen Basu, Mayank Gupta, Pratyaksha Rana et al.

We explore the potential of CNN-based models for gallbladder cancer (GBC) detection from ultrasound (USG) images as no prior study is known. USG is the most common diagnostic modality for GB diseases due to its low cost and accessibility. However, USG images are challenging to analyze due to low image quality, noise, and varying viewpoints due to the handheld nature of the sensor. Our exhaustive study of state-of-the-art (SOTA) image classification techniques for the problem reveals that they often fail to learn the salient GB region due to the presence of shadows in the USG images. SOTA object detection techniques also achieve low accuracy because of spurious textures due to noise or adjacent organs. We propose GBCNet to tackle the challenges in our problem. GBCNet first extracts the regions of interest (ROIs) by detecting the GB (and not the cancer), and then uses a new multi-scale, second-order pooling architecture specializing in classifying GBC. To effectively handle spurious textures, we propose a curriculum inspired by human visual acuity, which reduces the texture biases in GBCNet. Experimental results demonstrate that GBCNet significantly outperforms SOTA CNN models, as well as the expert radiologists. Our technical innovations are generic to other USG image analysis tasks as well. Hence, as a validation, we also show the efficacy of GBCNet in detecting breast cancer from USG images. Project page with source code, trained models, and data is available at https://gbc-iitd.github.io/gbcnet

Yatin Chaudhary, Pranav Rai, Matthias Schubert et al.

In this work, we combine the two paradigms: Federated Learning (FL) and Continual Learning (CL) for text classification task in cloud-edge continuum. The objective of Federated Continual Learning (FCL) is to improve deep learning models over life time at each client by (relevant and efficient) knowledge transfer without sharing data. Here, we address challenges in minimizing inter-client interference while knowledge sharing due to heterogeneous tasks across clients in FCL setup. In doing so, we propose a novel framework, Federated Selective Inter-client Transfer (FedSeIT) which selectively combines model parameters of foreign clients. To further maximize knowledge transfer, we assess domain overlap and select informative tasks from the sequence of historical tasks at each foreign client while preserving privacy. Evaluating against the baselines, we show improved performance, a gain of (average) 12.4\% in text classification over a sequence of tasks using five datasets from diverse domains. To the best of our knowledge, this is the first work that applies FCL to NLP.

Soumen Basu, Ashish Papanai, Mayank Gupta et al.

Automated detection of Gallbladder Cancer (GBC) from Ultrasound (US) images is an important problem, which has drawn increased interest from researchers. However, most of these works use difficult-to-acquire information such as bounding box annotations or additional US videos. In this paper, we focus on GBC detection using only image-level labels. Such annotation is usually available based on the diagnostic report of a patient, and do not require additional annotation effort from the physicians. However, our analysis reveals that it is difficult to train a standard image classification model for GBC detection. This is due to the low inter-class variance (a malignant region usually occupies only a small portion of a US image), high intra-class variance (due to the US sensor capturing a 2D slice of a 3D object leading to large viewpoint variations), and low training data availability. We posit that even when we have only the image level label, still formulating the problem as object detection (with bounding box output) helps a deep neural network (DNN) model focus on the relevant region of interest. Since no bounding box annotations is available for training, we pose the problem as weakly supervised object detection (WSOD). Motivated by the recent success of transformer models in object detection, we train one such model, DETR, using multi-instance-learning (MIL) with self-supervised instance selection to suit the WSOD task. Our proposed method demonstrates an improvement of AP and detection sensitivity over the SOTA transformer-based and CNN-based WSOD methods. Project page is at https://gbc-iitd.github.io/wsod-gbc

Arjun Choudhry, Pankaj Gupta, Inder Khatri et al.

Named Entity Recognition (NER) involves the identification and classification of named entities in unstructured text into predefined classes. NER in languages with limited resources, like French, is still an open problem due to the lack of large, robust, labelled datasets. In this paper, we propose a transformer-based NER approach for French using adversarial adaptation to similar domain or general corpora for improved feature extraction and better generalization. We evaluate our approach on three labelled datasets and show that our adaptation framework outperforms the corresponding non-adaptive models for various combinations of transformer models, source datasets and target corpora.

Arkajyoti Chakraborty, Inder Khatri, Arjun Choudhry et al.

Recent works on fake news detection have shown the efficacy of using emotions as a feature for improved performance. However, the cross-domain impact of emotion-guided features for fake news detection still remains an open problem. In this work, we propose an emotion-guided, domain-adaptive, multi-task approach for cross-domain fake news detection, proving the efficacy of emotion-guided models in cross-domain settings for various datasets.

Soumen Basu, Somanshu Singla, Mayank Gupta et al.

Rich temporal information and variations in viewpoints make video data an attractive choice for learning image representations using unsupervised contrastive learning (UCL) techniques. State-of-the-art (SOTA) contrastive learning techniques consider frames within a video as positives in the embedding space, whereas the frames from other videos are considered negatives. We observe that unlike multiple views of an object in natural scene videos, an Ultrasound (US) video captures different 2D slices of an organ. Hence, there is almost no similarity between the temporally distant frames of even the same US video. In this paper we propose to instead utilize such frames as hard negatives. We advocate mining both intra-video and cross-video negatives in a hardness-sensitive negative mining curriculum in a UCL framework to learn rich image representations. We deploy our framework to learn the representations of Gallbladder (GB) malignancy from US videos. We also construct the first large-scale US video dataset containing 64 videos and 15,800 frames for learning GB representations. We show that the standard ResNet50 backbone trained with our framework improves the accuracy of models pretrained with SOTA UCL techniques as well as supervised pretrained models on ImageNet for the GB malignancy detection task by 2-6%. We further validate the generalizability of our method on a publicly available lung US image dataset of COVID-19 pathologies and show an improvement of 1.5% compared to SOTA. Source code, dataset, and models are available at https://gbc-iitd.github.io/usucl.

Pankaj Gupta, Priya Mudgil, Niharika Dutta et al.

Background: Pancreatic cancer is one of the most aggressive cancers, with poor survival rates. Endoscopic ultrasound (EUS) is a key diagnostic modality, but its effectiveness is constrained by operator subjectivity. This study evaluates a Vision Transformer-based deep learning segmentation model for pancreatic tumors. Methods: A segmentation model using the USFM framework with a Vision Transformer backbone was trained and validated with 17,367 EUS images (from two public datasets) in 5-fold cross-validation. The model was tested on an independent dataset of 350 EUS images from another public dataset, manually segmented by radiologists. Preprocessing included grayscale conversion, cropping, and resizing to 512x512 pixels. Metrics included Dice similarity coefficient (DSC), intersection over union (IoU), sensitivity, specificity, and accuracy. Results: In 5-fold cross-validation, the model achieved a mean DSC of 0.651 +/- 0.738, IoU of 0.579 +/- 0.658, sensitivity of 69.8%, specificity of 98.8%, and accuracy of 97.5%. For the external validation set, the model achieved a DSC of 0.657 (95% CI: 0.634-0.769), IoU of 0.614 (95% CI: 0.590-0.689), sensitivity of 71.8%, and specificity of 97.7%. Results were consistent, but 9.7% of cases exhibited erroneous multiple predictions. Conclusions: The Vision Transformer-based model demonstrated strong performance for pancreatic tumor segmentation in EUS images. However, dataset heterogeneity and limited external validation highlight the need for further refinement, standardization, and prospective studies.

Arjun Choudhry, Inder Khatri, Pankaj Gupta et al.

Named Entity Recognition (NER) is the task of identifying and classifying named entities in large-scale texts into predefined classes. NER in French and other relatively limited-resource languages cannot always benefit from approaches proposed for languages like English due to a dearth of large, robust datasets. In this paper, we present our work that aims to mitigate the effects of this dearth of large, labeled datasets. We propose a Transformer-based NER approach for French, using adversarial adaptation to similar domain or general corpora to improve feature extraction and enable better generalization. Our approach allows learning better features using large-scale unlabeled corpora from the same domain or mixed domains to introduce more variations during training and reduce overfitting. Experimental results on three labeled datasets show that our adaptation framework outperforms the corresponding non-adaptive models for various combinations of Transformer models, source datasets, and target corpora. We also show that adversarial adaptation to large-scale unlabeled corpora can help mitigate the performance dip incurred on using Transformer models pre-trained on smaller corpora.

Chetan Madan, Mayuna Gupta, Soumen Basu et al.

We focus on the problem of Gallbladder Cancer (GBC) detection from Ultrasound (US) images. The problem presents unique challenges to modern Deep Neural Network (DNN) techniques due to low image quality arising from noise, textures, and viewpoint variations. Tackling such challenges would necessitate precise localization performance by the DNN to identify the discerning features for the downstream malignancy prediction. While several techniques have been proposed in the recent years for the problem, all of these methods employ complex custom architectures. Inspired by the success of foundational models for natural image tasks, along with the use of adapters to fine-tune such models for the custom tasks, we investigate the merit of one such design, ViT-Adapter, for the GBC detection problem. We observe that ViT-Adapter relies predominantly on a primitive CNN-based spatial prior module to inject the localization information via cross-attention, which is inefficient for our problem due to the small pathology sizes, and variability in their appearances due to non-regular structure of the malignancy. In response, we propose, LQ-Adapter, a modified Adapter design for ViT, which improves localization information by leveraging learnable content queries over the basic spatial prior module. Our method surpasses existing approaches, enhancing the mean IoU (mIoU) scores by 5.4%, 5.8%, and 2.7% over ViT-Adapters, DINO, and FocalNet-DINO, respectively on the US image-based GBC detection dataset, and establishing a new state-of-the-art (SOTA). Additionally, we validate the applicability and effectiveness of LQ-Adapter on the Kvasir-Seg dataset for polyp detection from colonoscopy images. Superior performance of our design on this problem as well showcases its capability to handle diverse medical imaging tasks across different datasets. Code is released at https://github.com/ChetanMadan/LQ-Adapter

Pankaj Gupta, Kartik Bose

Large language models (LLMs) show promise in radiology but their deployment is limited by computational requirements that preclude use in resource-constrained clinical environments. We investigate whether small language models (SLMs) of 3-4 billion parameters can achieve strong multi-task radiology performance through LoRA fine-tuning, enabling deployment on consumer-grade CPUs. We train Qwen2.5-3B-Instruct and Qwen3-4B on 162K samples spanning 9 radiology tasks - RADS classification across 10 systems, impression generation, temporal comparison, radiology NLI, NER, abnormality detection, N/M staging, and radiology Q&A - compiled from 12 public datasets. Both models are evaluated on up to 500 held-out test samples per task with standardized metrics. Our key findings are: (1) LoRA fine-tuning dramatically improves performance over zero-shot baselines (RADS accuracy +53%, NLI +60%, N-staging +89%); (2) the two models exhibit complementary strengths - Qwen2.5 excels at structured generation tasks while Qwen3 dominates extractive tasks; (3) a task-outed oracle ensemble combining both models achieves the best performance across all tasks; (4) few-shot prompting with fine-tuned models hurts performance, demonstrating that LoRA adaptation is more effective than in-context learning for specialized domains; and (5) models can be quantized to GGUF format (~1.8-2.4GB) for CPU deployment at 4-8 tokens/second on consumer hardware. Our work demonstrates that small, efficiently fine-tuned models - which we collectively call RadLite - can serve as practical multi-task radiology AI assistants deployable entirely on consumer hardware without GPU requirements.

Chetan Madan, Aarjav Satia, Soumen Basu et al.

Masked Autoencoders (MAEs) have emerged as a dominant strategy for self-supervised representation learning in natural images, where models are pre-trained to reconstruct masked patches with a pixel-wise mean squared error (MSE) between original and reconstructed RGB values as the loss. We observe that MSE encourages blurred image re-construction, but still works for natural images as it preserves dominant edges. However, in medical imaging, when the texture cues are more important for classification of a visual abnormality, the strategy fails. Taking inspiration from Gray Level Co-occurrence Matrix (GLCM) feature in Radiomics studies, we propose a novel MAE based pre-training framework, GLCM-MAE, using reconstruction loss based on matching GLCM. GLCM captures intensity and spatial relationships in an image, hence proposed loss helps preserve morphological features. Further, we propose a novel formulation to convert matching GLCM matrices into a differentiable loss function. We demonstrate that unsupervised pre-training on medical images with the proposed GLCM loss improves representations for downstream tasks. GLCM-MAE outperforms the current state-of-the-art across four tasks - gallbladder cancer detection from ultrasound images by 2.1%, breast cancer detection from ultrasound by 3.1%, pneumonia detection from x-rays by 0.5%, and COVID detection from CT by 0.6%. Source code and pre-trained models are available at: https://github.com/ChetanMadan/GLCM-MAE.

Shivika, Kartik Bose, Pankaj Gupta

Vision-language models trained with contrastive learning on paired medical images and reports show strong zero-shot diagnostic capabilities, yet the effect of training batch composition on learned representations remains unexplored for 3D medical imaging. We reproduce Merlin, a dual-encoder model that aligns 3D abdominal CT volumes with radiology reports using symmetric InfoNCE loss, achieving a zero-shot macro F1 of 74.45% across 30 findings (original: 73.00%). We then investigate two axes of variation. First, we control the normal-to-abnormal ratio within training batches at 25:75, 50:50, and 75:25 using section-level balanced sampling on the full dataset. All three configurations underperform the unbalanced baseline by 2.4 to 2.8 points, with 75:25 achieving the best result (72.02%) among balanced variants. Second, we conduct data scaling ablations on a 4,362-study subset, training with 20%, 40%, and 100% of the data. Performance scales sub-linearly from 65.26% to 71.88%, with individual findings varying dramatically in data sensitivity. Enforcing 50:50 balanced sampling on the same subset further degrades performance to 68.01%, confirming that explicit class balancing hurts regardless of dataset or balancing granularity. Our results indicate that the stochastic diversity of random sampling, combined with Merlin's alternating batching over anatomical subsections, provides more effective regularization than engineered class ratios at the small batch sizes required by 3D medical volumes.

Soumen Basu, Mayuna Gupta, Chetan Madan et al.

In recent years, automated Gallbladder Cancer (GBC) detection has gained the attention of researchers. Current state-of-the-art (SOTA) methodologies relying on ultrasound sonography (US) images exhibit limited generalization, emphasizing the need for transformative approaches. We observe that individual US frames may lack sufficient information to capture disease manifestation. This study advocates for a paradigm shift towards video-based GBC detection, leveraging the inherent advantages of spatiotemporal representations. Employing the Masked Autoencoder (MAE) for representation learning, we address shortcomings in conventional image-based methods. We propose a novel design called FocusMAE to systematically bias the selection of masking tokens from high-information regions, fostering a more refined representation of malignancy. Additionally, we contribute the most extensive US video dataset for GBC detection. We also note that, this is the first study on US video-based GBC detection. We validate the proposed methods on the curated dataset, and report a new state-of-the-art (SOTA) accuracy of 96.4% for the GBC detection problem, against an accuracy of 84% by current Image-based SOTA - GBCNet, and RadFormer, and 94.7% by Video-based SOTA - AdaMAE. We further demonstrate the generality of the proposed FocusMAE on a public CT-based Covid detection dataset, reporting an improvement in accuracy by 3.3% over current baselines. The source code and pretrained models are available at: https://gbc-iitd.github.io/focusmae

Pankaj Gupta, Yatin Chaudhary, Hinrich Schütze

Though word embeddings and topics are complementary representations, several past works have only used pretrained word embeddings in (neural) topic modeling to address data sparsity in short-text or small collection of documents. This work presents a novel neural topic modeling framework using multi-view embedding spaces: (1) pretrained topic-embeddings, and (2) pretrained word-embeddings (context insensitive from Glove and context-sensitive from BERT models) jointly from one or many sources to improve topic quality and better deal with polysemy. In doing so, we first build respective pools of pretrained topic (i.e., TopicPool) and word embeddings (i.e., WordPool). We then identify one or more relevant source domain(s) and transfer knowledge to guide meaningful learning in the sparse target domain. Within neural topic modeling, we quantify the quality of topics and document representations via generalization (perplexity), interpretability (topic coherence) and information retrieval (IR) using short-text, long-text, small and large document collections from news and medical domains. Introducing the multi-source multi-view embedding spaces, we have shown state-of-the-art neural topic modeling using 6 source (high-resource) and 5 target (low-resource) corpora.

Yatin Chaudhary, Pankaj Gupta, Khushbu Saxena et al.

Prior research notes that BERT's computational cost grows quadratically with sequence length thus leading to longer training times, higher GPU memory constraints and carbon emissions. While recent work seeks to address these scalability issues at pre-training, these issues are also prominent in fine-tuning especially for long sequence tasks like document classification. Our work thus focuses on optimizing the computational cost of fine-tuning for document classification. We achieve this by complementary learning of both topic and language models in a unified framework, named TopicBERT. This significantly reduces the number of self-attention operations - a main performance bottleneck. Consequently, our model achieves a 1.4x ($\sim40\%$) speedup with $\sim40\%$ reduction in $CO_2$ emission while retaining $99.9\%$ performance over 5 datasets.

Pankaj Gupta, Yatin Chaudhary, Thomas Runkler et al.

Lifelong learning has recently attracted attention in building machine learning systems that continually accumulate and transfer knowledge to help future learning. Unsupervised topic modeling has been popularly used to discover topics from document collections. However, the application of topic modeling is challenging due to data sparsity, e.g., in a small collection of (short) documents and thus, generate incoherent topics and sub-optimal document representations. To address the problem, we propose a lifelong learning framework for neural topic modeling that can continuously process streams of document collections, accumulate topics and guide future topic modeling tasks by knowledge transfer from several sources to better deal with the sparse data. In the lifelong process, we particularly investigate jointly: (1) sharing generative homologies (latent topics) over lifetime to transfer prior knowledge, and (2) minimizing catastrophic forgetting to retain the past learning via novel selective data augmentation, co-training and topic regularization approaches. Given a stream of document collections, we apply the proposed Lifelong Neural Topic Modeling (LNTM) framework in modeling three sparse document collections as future tasks and demonstrate improved performance quantified by perplexity, topic coherence and information retrieval task.

Yatin Chaudhary, Hinrich Schütze, Pankaj Gupta

Marrying topic models and language models exposes language understanding to a broader source of document-level context beyond sentences via topics. While introducing topical semantics in language models, existing approaches incorporate latent document topic proportions and ignore topical discourse in sentences of the document. This work extends the line of research by additionally introducing an explainable topic representation in language understanding, obtained from a set of key terms correspondingly for each latent topic of the proportion. Moreover, we retain sentence-topic associations along with document-topic association by modeling topical discourse for every sentence in the document. We present a novel neural composite language model that exploits both the latent and explainable topics along with topical discourse at sentence-level in a joint learning framework of topic and language models. Experiments over a range of tasks such as language modeling, word sense disambiguation, document classification, retrieval and text generation demonstrate ability of the proposed model in improving language understanding.

Usama Yaseen, Pankaj Gupta, Hinrich Schütze

Named Entity Recognition (NER) and Relation Extraction (RE) are essential tools in distilling knowledge from biomedical literature. This paper presents our findings from participating in BioNLP Shared Tasks 2019. We addressed Named Entity Recognition including nested entities extraction, Entity Normalization and Relation Extraction. Our proposed approach of Named Entities can be generalized to different languages and we have shown it's effectiveness for English and Spanish text. We investigated linguistic features, hybrid loss including ranking and Conditional Random Fields (CRF), multi-task objective and token-level ensembling strategy to improve NER. We employed dictionary based fuzzy and semantic search to perform Entity Normalization. Finally, our RE system employed Support Vector Machine (SVM) with linguistic features. Our NER submission (team:MIC-CIS) ranked first in BB-2019 norm+NER task with standard error rate (SER) of 0.7159 and showed competitive performance on PharmaCo NER task with F1-score of 0.8662. Our RE system ranked first in the SeeDev-binary Relation Extraction Task with F1-score of 0.3738.

Yatin Chaudhary, Pankaj Gupta, Hinrich Schütze

This paper presents our system details and results of participation in the RDoC Tasks of BioNLP-OST 2019. Research Domain Criteria (RDoC) construct is a multi-dimensional and broad framework to describe mental health disorders by combining knowledge from genomics to behaviour. Non-availability of RDoC labelled dataset and tedious labelling process hinders the use of RDoC framework to reach its full potential in Biomedical research community and Healthcare industry. Therefore, Task-1 aims at retrieval and ranking of PubMed abstracts relevant to a given RDoC construct and Task-2 aims at extraction of the most relevant sentence from a given PubMed abstract. We investigate (1) attention based supervised neural topic model and SVM for retrieval and ranking of PubMed abstracts and, further utilize BM25 and other relevance measures for re-ranking, (2) supervised and unsupervised sentence ranking models utilizing multi-view representations comprising of query-aware attention-based sentence representation (QAR), bag-of-words (BoW) and TF-IDF. Our best systems achieved 1st rank and scored 0.86 mean average precision (mAP) and 0.58 macro average accuracy (MAA) in Task-1 and Task-2 respectively.

Yatin Chaudhary, Pankaj Gupta, Thomas Runkler

Topic models such as LDA, DocNADE, iDocNADEe have been popular in document analysis. However, the traditional topic models have several limitations including: (1) Bag-of-words (BoW) assumption, where they ignore word ordering, (2) Data sparsity, where the application of topic models is challenging due to limited word co-occurrences, leading to incoherent topics and (3) No Continuous Learning framework for topic learning in lifelong fashion, exploiting historical knowledge (or latent topics) and minimizing catastrophic forgetting. This thesis focuses on addressing the above challenges within neural topic modeling framework. We propose: (1) Contextualized topic model that combines a topic and a language model and introduces linguistic structures (such as word ordering, syntactic and semantic features, etc.) in topic modeling, (2) A novel lifelong learning mechanism into neural topic modeling framework to demonstrate continuous learning in sequential document collections and minimizing catastrophic forgetting. Additionally, we perform a selective data augmentation to alleviate the need for complete historical corpora during data hallucination or replay.

Pankaj Gupta, Yatin Chaudhary, Hinrich Schütze

Though word embeddings and topics are complementary representations, several past works have only used pre-trained word embeddings in (neural) topic modeling to address data sparsity problem in short text or small collection of documents. However, no prior work has employed (pre-trained latent) topics in transfer learning paradigm. In this paper, we propose an approach to (1) perform knowledge transfer using latent topics obtained from a large source corpus, and (2) jointly transfer knowledge via the two representations (or views) in neural topic modeling to improve topic quality, better deal with polysemy and data sparsity issues in a target corpus. In doing so, we first accumulate topics and word representations from one or many source corpora to build a pool of topics and word vectors. Then, we identify one or multiple relevant source domain(s) and take advantage of corresponding topics and word features via the respective pools to guide meaningful learning in the sparse target domain. We quantify the quality of topic and document representations via generalization (perplexity), interpretability (topic coherence) and information retrieval (IR) using short-text, long-text, small and large document collections from news and medical domains. We have demonstrated the state-of-the-art results on topic modeling with the proposed framework.

Pankaj Gupta, Khushbu Saxena, Usama Yaseen et al.

This paper describes our system (MIC-CIS) details and results of participation in the fine-grained propaganda detection shared task 2019. To address the tasks of sentence (SLC) and fragment level (FLC) propaganda detection, we explore different neural architectures (e.g., CNN, LSTM-CRF and BERT) and extract linguistic (e.g., part-of-speech, named entity, readability, sentiment, emotion, etc.), layout and topical features. Specifically, we have designed multi-granularity and multi-tasking neural architectures to jointly perform both the sentence and fragment level propaganda detection. Additionally, we investigate different ensemble schemes such as majority-voting, relax-voting, etc. to boost overall system performance. Compared to the other participating systems, our submissions are ranked 3rd and 4th in FLC and SLC tasks, respectively.

Pankaj Gupta, Subburam Rajaram, Hinrich Schütze et al.

Past work in relation extraction mostly focuses on binary relation between entity pairs within single sentence. Recently, the NLP community has gained interest in relation extraction in entity pairs spanning multiple sentences. In this paper, we propose a novel architecture for this task: inter-sentential dependency-based neural networks (iDepNN). iDepNN models the shortest and augmented dependency paths via recurrent and recursive neural networks to extract relationships within (intra-) and across (inter-) sentence boundaries. Compared to SVM and neural network baselines, iDepNN is more robust to false positives in relationships spanning sentences. We evaluate our models on four datasets from newswire (MUC6) and medical (BioNLP shared task) domains that achieve state-of-the-art performance and show a better balance in precision and recall for inter-sentential relationships. We perform better than 11 teams participating in the BioNLP shared task 2016 and achieve a gain of 5.2% (0.587 vs 0.558) in F1 over the winning team. We also release the crosssentence annotations for MUC6.

Pankaj Gupta, Yatin Chaudhary, Florian Buettner et al.

We address two challenges of probabilistic topic modelling in order to better estimate the probability of a word in a given context, i.e., P(word|context): (1) No Language Structure in Context: Probabilistic topic models ignore word order by summarizing a given context as a "bag-of-word" and consequently the semantics of words in the context is lost. The LSTM-LM learns a vector-space representation of each word by accounting for word order in local collocation patterns and models complex characteristics of language (e.g., syntax and semantics), while the TM simultaneously learns a latent representation from the entire document and discovers the underlying thematic structure. We unite two complementary paradigms of learning the meaning of word occurrences by combining a TM (e.g., DocNADE) and a LM in a unified probabilistic framework, named as ctx-DocNADE. (2) Limited Context and/or Smaller training corpus of documents: In settings with a small number of word occurrences (i.e., lack of context) in short text or data sparsity in a corpus of few documents, the application of TMs is challenging. We address this challenge by incorporating external knowledge into neural autoregressive topic models via a language modelling approach: we use word embeddings as input of a LSTM-LM with the aim to improve the word-topic mapping on a smaller and/or short-text corpus. The proposed DocNADE extension is named as ctx-DocNADEe. We present novel neural autoregressive topic model variants coupled with neural LMs and embeddings priors that consistently outperform state-of-the-art generative TMs in terms of generalization (perplexity), interpretability (topic coherence) and applicability (retrieval and classification) over 6 long-text and 8 short-text datasets from diverse domains.

Pankaj Gupta, Yatin Chaudhary, Florian Buettner et al.

We address two challenges in topic models: (1) Context information around words helps in determining their actual meaning, e.g., "networks" used in the contexts "artificial neural networks" vs. "biological neuron networks". Generative topic models infer topic-word distributions, taking no or only little context into account. Here, we extend a neural autoregressive topic model to exploit the full context information around words in a document in a language modeling fashion. The proposed model is named as iDocNADE. (2) Due to the small number of word occurrences (i.e., lack of context) in short text and data sparsity in a corpus of few documents, the application of topic models is challenging on such texts. Therefore, we propose a simple and efficient way of incorporating external knowledge into neural autoregressive topic models: we use embeddings as a distributional prior. The proposed variants are named as DocNADEe and iDocNADEe. We present novel neural autoregressive topic model variants that consistently outperform state-of-the-art generative topic models in terms of generalization, interpretability (topic coherence) and applicability (retrieval and classification) over 7 long-text and 8 short-text datasets from diverse domains.

Pankaj Gupta, Florian Buettner, Hinrich Schütze

Context information around words helps in determining their actual meaning, for example "networks" used in contexts of artificial neural networks or biological neuron networks. Generative topic models infer topic-word distributions, taking no or only little context into account. Here, we extend a neural autoregressive topic model to exploit the full context information around words in a document in a language modeling fashion. This results in an improved performance in terms of generalization, interpretability and applicability. We apply our modeling approach to seven data sets from various domains and demonstrate that our approach consistently outperforms stateof-the-art generative topic models. With the learned representations, we show on an average a gain of 9.6% (0.57 Vs 0.52) in precision at retrieval fraction 0.02 and 7.2% (0.582 Vs 0.543) in F1 for text categorization.

Pankaj Gupta, Hinrich Schütze

Recurrent neural networks (RNNs) are temporal networks and cumulative in nature that have shown promising results in various natural language processing tasks. Despite their success, it still remains a challenge to understand their hidden behavior. In this work, we analyze and interpret the cumulative nature of RNN via a proposed technique named as Layer-wIse-Semantic-Accumulation (LISA) for explaining decisions and detecting the most likely (i.e., saliency) patterns that the network relies on while decision making. We demonstrate (1) LISA: "How an RNN accumulates or builds semantics during its sequential processing for a given text example and expected response" (2) Example2pattern: "How the saliency patterns look like for each category in the data according to the network in decision making". We analyse the sensitiveness of RNNs about different inputs to check the increase or decrease in prediction scores and further extract the saliency patterns learned by the network. We employ two relation classification datasets: SemEval 10 Task 8 and TAC KBP Slot Filling to explain RNN predictions via the LISA and example2pattern.

Pankaj Gupta, Bernt Andrassy, Hinrich Schütze

The goal of our industrial ticketing system is to retrieve a relevant solution for an input query, by matching with historical tickets stored in knowledge base. A query is comprised of subject and description, while a historical ticket consists of subject, description and solution. To retrieve a relevant solution, we use textual similarity paradigm to learn similarity in the query and historical tickets. The task is challenging due to significant term mismatch in the query and ticket pairs of asymmetric lengths, where subject is a short text but description and solution are multi-sentence texts. We present a novel Replicated Siamese LSTM model to learn similarity in asymmetric text pairs, that gives 22% and 7% gain (Accuracy@10) for retrieval task, respectively over unsupervised and supervised baselines. We also show that the topic and distributed semantic features for short and long texts improved both similarity learning and retrieval.

Pankaj Gupta, Benjamin Roth, Hinrich Schütze

Semi-supervised bootstrapping techniques for relationship extraction from text iteratively expand a set of initial seed instances. Due to the lack of labeled data, a key challenge in bootstrapping is semantic drift: if a false positive instance is added during an iteration, then all following iterations are contaminated. We introduce BREX, a new bootstrapping method that protects against such contamination by highly effective confidence assessment. This is achieved by using entity and template seeds jointly (as opposed to just one as in previous work), by expanding entities and templates in parallel and in a mutually constraining fashion in each iteration and by introducing higherquality similarity measures for templates. Experimental results show that BREX achieves an F1 that is 0.13 (0.87 vs. 0.74) better than the state of the art for four relationships.

Pankaj Gupta, Subburam Rajaram, Hinrich Schütze et al.

Dynamic topic modeling facilitates the identification of topical trends over time in temporal collections of unstructured documents. We introduce a novel unsupervised neural dynamic topic model named as Recurrent Neural Network-Replicated Softmax Model (RNNRSM), where the discovered topics at each time influence the topic discovery in the subsequent time steps. We account for the temporal ordering of documents by explicitly modeling a joint distribution of latent topical dependencies over time, using distributional estimators with temporal recurrent connections. Applying RNN-RSM to 19 years of articles on NLP research, we demonstrate that compared to state-of-the art topic models, RNNRSM shows better generalization, topic interpretation, evolution and trends. We also introduce a metric (named as SPAN) to quantify the capability of dynamic topic model to capture word evolution in topics over time.

Drew Linsley, Sven Eberhardt, Tarun Sharma et al.

Although Deep Convolutional Networks (DCNs) are approaching the accuracy of human observers at object recognition, it is unknown whether they leverage similar visual representations to achieve this performance. To address this, we introduce Clicktionary, a web-based game for identifying visual features used by human observers during object recognition. Importance maps derived from the game are consistent across participants and uncorrelated with image saliency measures. These results suggest that Clicktionary identifies image regions that are meaningful and diagnostic for object recognition but different than those driving eye movements. Surprisingly, Clicktionary importance maps are only weakly correlated with relevance maps derived from DCNs trained for object recognition. Our study demonstrates that the narrowing gap between the object recognition accuracy of human observers and DCNs obscures distinct visual strategies used by each to achieve this performance.

Ngoc Thang Vu, Heike Adel, Pankaj Gupta et al.

This paper investigates two different neural architectures for the task of relation classification: convolutional neural networks and recurrent neural networks. For both models, we demonstrate the effect of different architectural choices. We present a new context representation for convolutional neural networks for relation classification (extended middle context). Furthermore, we propose connectionist bi-directional recurrent neural networks and introduce ranking loss for their optimization. Finally, we show that combining convolutional and recurrent neural networks using a simple voting scheme is accurate enough to improve results. Our neural models achieve state-of-the-art results on the SemEval 2010 relation classification task.