Scholars@Duke
Dhanajit Brahma

Scholarly Works

SAGE: A Novelty Gate for Efficient Memory Evolution in Agentic LLMs

Preprint · May 28, 2026 Agentic LLMs must continuously decide whether newly extracted facts should be added, merged with existing memories, or ignored, yet prior work has focused more on retrieval and storage than on principled write-side control. We frame memory evolution as a n ... Link to item Cite

Calibrating Model-Based Evaluation Metrics for Summarization

Conference · April 18, 2026 Recent advances in summary evaluation are based on model-based metrics to assess quality dimensions, such as completeness, conciseness, and faithfulness. However, these methods often require large language models, and predicted scores are frequently miscal ... Link to item Cite

Coupling Generative Modeling and an Autoencoder with the Causal Bridge

Conference · September 29, 2025 We consider inferring the causal effect of a treatment (intervention) on an outcome of interest in situations where there is potentially an unobserved confounder influencing both the treatment and the outcome. This is achievable by assuming access to two s ... Link to item Cite

Text-Based Fine-Grained Emotion Prediction

Journal article IEEE Transactions on Affective Computing · April 1, 2024 Text-based emotion prediction is an important task in the field of affective computing. Most prior work has been restricted to predicting emotions corresponding to a few high-level emotion classes. This paper explores and experiments with various technique ... Full text Cite

A Probabilistic Framework for Lifelong Test-Time Adaptation

Conference Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition · January 1, 2023 Test-time adaptation (TTA) is the problem of updating a pre-trained source model at inference time given test input(s) from a different target domain. Most existing TTA approaches assume the setting in which the target domain is stationary, i.e., all the t ... Full text Cite

Fine-Grained Emotion Prediction by Modeling Emotion Definitions

Conference 2021 9th International Conference on Affective Computing and Intelligent Interaction Acii 2021 · January 1, 2021 In this paper, we propose a new framework for fine-grained emotion prediction in the text through emotion definition modeling. Our approach involves a multi-task learning framework that models definitions of emotions as an auxiliary task while being traine ... Full text Cite

Deep attentive ranking networks for learning to order sentences

Journal article Aaai 2020 34th Aaai Conference on Artificial Intelligence · January 1, 2020 We present an attention-based ranking framework for learning to order sentences given a paragraph. Our framework is built on a bidirectional sentence encoder and a self-attention based transformer network to obtain an input order invariant representation o ... Cite

Meta-learning for generalized zero-shot learning

Journal article Aaai 2020 34th Aaai Conference on Artificial Intelligence · January 1, 2020 Learning to classify unseen class samples at test time is popularly referred to as zero-shot learning (ZSL). If test samples can be from training (seen) as well as unseen classes, it is a more challenging problem due to the existence of strong bias towards ... Cite

An Integrated Approach for Identification of Functionally Similar MicroRNAs in Colorectal Cancer.

Journal article IEEE/ACM transactions on computational biology and bioinformatics · January 2019 Colorectal cancer (CRC) is one of the most prevalent cancers around the globe. However, the molecular reasons for pathogenesis of CRC are still poorly understood. Recently, the role of microRNAs or miRNAs in the initiation and progression of CRC has been s ... Full text Cite