Around April 2025, I started working on a paper for ICLR. The plan was to collaborate (equally) with one of my PhD supervisor’s students, but as time went on, I took on most of the responsibility and ended up writing the entire paper + coding all the main results and ablations. The other student ran some baselines, but the results had mistakes. So I had to re-implement and correct the baselines. In the final version, everything including writing, […]

Ver mais

Obtaining labelled ECG data for developing supervised models is challenging. Contrastive learning (CL) has emerged as a promising pretraining approach that enables effective transfer learning with limited labelled data. However, existing CL frameworks either focus solely on global context or fail to exploit ECG-specific characteristics. Furthermore, these methods rely on hard contrastive targets, which may not adequately capture the continuous nature of feature similarity in ECG signals. In this paper, we propose Beat-SSL, a contrastive learning framework that […]

Ver mais

Specifications for decentralized learning on resource-constrained edge devices require algorithms that are communication-efficient, robust to data corruption, and lightweight in memory usage. While state-of-the-art gossip-based methods satisfy the first requirement, achieving robustness remains challenging. Asynchronous decentralized ADMM-based methods have been explored for estimating the median, a statistical centrality measure that is notoriously more robust than the mean. However, existing approaches require memory that scales with node degree, making them impractical when memory is limited. In this paper, we […]

Ver mais

Photoacoustic computed tomography (PACT) is a promising imaging modality that combines the advantages of optical contrast with ultrasound detection. Utilizing ultrasound transducers with larger surface areas can improve detection sensitivity. However, when computationally efficient analytic reconstruction methods that neglect the spatial impulse responses (SIRs) of the transducer are employed, the spatial resolution of the reconstructed images will be compromised. Although optimization-based reconstruction methods can explicitly account for SIR effects, their computational cost is generally high, particularly in three-dimensional […]

Ver mais

Key Highlights: There’s no denying that the BCCI is the biggest cricket board in the world, with reported profit projections of nearly ₹6,700 crore for FY2025–26. The BCCI’s projections also include massive contributions from the Indian Premier League (IPL). If you live in India or around the world and are familiar with cricket, you must be aware of the IPL. It is the biggest franchise cricket league in terms of viewership, valuation, and star power, followed by South […]

Ver mais

Language models are revolutionizing the biochemistry domain, assisting scientists in drug design and chemical synthesis with high efficiency. Yet current approaches struggle between small language models prone to hallucination and limited knowledge retention, and large cloud-based language models plagued by privacy risks and high inference costs. To bridge this gap, we introduce ChemCRAFT, a novel framework leveraging agentic reinforcement learning to decouple chemical reasoning from knowledge storage. Instead of forcing the model to memorize vast chemical data, our […]

Ver mais

Brain tumor segmentation is a critical process in medical imaging, enabling early diagnosis and effective treatment planning. Traditional manual segmentation is time-consuming, prone to inter-observer variability, and requires skilled radiologists. While deep learning has shown promising results in medical image analysis, existing research often faces generalization challenges, suffers from class imbalance, or lacks effective architectures specifically designed for 3D MRI data. Many existing works focus on 2D segmentation, limiting spatial contextual understanding, or employ models with suboptimal accuracy […]

Ver mais