Srirangaraj Setlur

Child handwriting enhancement poses unique challenges distinct from other document restoration tasks. Unlike the global degradation observed in historical manuscripts, children’s handwriting exhibits locally inconsistent faint strokes, misalignment with the writing line, and behavioral artifacts such as scribbles or partial erasures. When scanned, these irregularities are compounded by background noise, illumination variation, and scanner-induced distortions, deforming vital stroke information. Existing document restoration models, optimized for high-contrast ink or uniform degradations, tend to erase these subtle pencil traces, severely limiting their utility for downstream child handwriting analysis. To address these issues, we present TraceNet, an application-driven restoration network designed specifically for faint and irregular child handwriting. TraceNet integrates Swin Transformer-based self-attention within a single deterministic encoder-decoder restoration framework. Temporal diffusion embeddings are used as a conditioning signal to adapt restoration strength according to degradation severity, enabling selective recovery of faint strokes and line removal without over-smoothing. Evaluations on a private corpus of child handwriting show that TraceNet achieves strong performance across image-quality and downstream HTR metrics over state-of-the-art baselines. An additional evaluation on the DIBCO dataset further shows that the same framework is effective for historical document enhancement. Remarkably, the model maintains high restoration quality even under few-shot fine-tuning with only a small set of annotated samples, highlighting its efficiency and adaptability for real-world educational, archival, and behavioral handwriting analysis tasks.

Artificial Intelligence (AI) research increasingly relies on large, interdisciplinary teams to address complex societal challenges. However, coordinating these team science efforts presents challenges such as communication barriers and operational inefficiencies. This paper analyzes insights from project managers (PMs) across 25 U.S. National AI Research Institutes, examining how PMs foster collective intelligence, strengthen group dynamics, improve operational efficiency, and cultivate cross-disciplinary relationships. This work demonstrates that PMs can play a pivotal role in accelerating research outcomes, with recommendations for funding agencies, academic institutions, and research organizations on how to integrate and support project management to increase the impact of scientific teams. Although grounded in the context of AI research, these findings have broader relevance for large-scale, interdisciplinary research initiatives across the sciences.

Postoperative complications remain a critical concern in clinical practice, adversely affecting patient outcomes and contributing to rising healthcare costs. We present MIRACLE, a deep learning architecture for prediction of risk of postoperative complications in lung cancer surgery by integrating preoperative clinical and radiological data. MIRACLE employs a hyperspherical embedding space fusion of heterogeneous inputs, enabling the extraction of robust, discriminative features from both structured clinical records and high-dimensional radiological images. To enhance transparency of prediction and clinical utility, we incorporate an interventional deep learning module in MIRACLE, that not only refines predictions but also provides interpretable and actionable insights, allowing domain experts to interactively adjust recommendations based on clinical expertise. We validate our approach on POC-L, a real-world dataset comprising 3,094 lung cancer patients who underwent surgery at Roswell Park Comprehensive Cancer Center. Our results demonstrate that MIRACLE outperforms various traditional machine learning models and contemporary large language models (LLM) variants alone, for personalized and explainable postoperative risk management.

Signaling theory elaborates how communicative patterns and signals, such as facial expressions, have facilitated interactions hence helped social animals survive. Signaling theory suggests that those sending clearer signals, through the face or otherwise, may be more trustworthy communicators. This would apply even to negative emotional signals such as disgust. In contrast, the overgeneralization hypothesis suggests a clear negative facial expression might overgeneralize to have a deleterious effect on judgments of someone’s trustworthiness. We tested this in a study where 64 males and 70 female judges viewed 12 video clips, featuring 6 men and 6 women showing either a macro expression (M = 2.66s duration) or a micro expression (M = 0.34s duration) of disgust, as they spontaneously reacted to a bad-tasting drink. The judges then rated the participants on how trustworthy, authentic believable, and likeable they appeared. A repeated measures MANOVA found all macro expression signal givers to be perceived significantly more trustworthy, authentic, believable, and likeable compared with micro expression signal givers. A significant interaction between length of expression and gender showed when expressing macro expressions, female expressers were perceived equally positive to males, yet when expressing micro expressions, female expressers were perceived significantly less trustworthy, authentic, believable, and likeable than males. This study supports a signaling theory explanation and adds further insight into the role clarity of signals has in signaling theory.

Dyslexia and dysgraphia are two specific learning disabilities (SLDs) that are prevalent among children. To minimize the negative impact these SLDs have on a child’s academic and social-emotional development, it is crucial to identify dyslexia and dysgraphia at an early age, enabling timely and effective intervention. The first step in this process is screening, which helps determine if a child requires further instruction or a more in-depth assessment. Current screening tools are expensive, require additional administration time beyond regular classroom activities, and are designed to screen exclusively for one condition, not for both dyslexia and dysgraphia, which often share some common behavioral characteristics. Most dyslexia screeners focus on speech and oral tasks and exclude writing activities. However, analyzing children’s writing samples for behavioral signs of dyslexia and dysgraphia can offer valuable insights into the screening process, which can be time-consuming. As a solution, we propose a co-designed framework for building artificial intelligence (AI) tools that could boost the efficiency of screening and aid practitioners such as speech-language pathologists (SLPs), occupational therapists, general educators, and special educators by simplifying their tasks. This paper reviews current screening methods employed by practitioners, the use of AI-based systems in identifying dyslexia and dysgraphia, and the handwriting datasets available to train such systems. The paper also outlines a framework for developing an AI-integrated screening tool that can identify writing-based behavioral indicators of dyslexia and dysgraphia in children’s handwriting. This framework can be used in conjunction with current screening tools like the Dysgraphia and Dyslexia Behavioral Indicator Checklist (DDBIC). The paper also proposes a methodology for collecting children’s offline and online handwriting samples to build a valuable dataset for developing AI solutions. The proposed framework and data collection methodology are co-designed with SLPs, occupational therapists (OTs), special educators, and general educators to ensure the tool can provide explainable, actionable information that would be invaluable in a practical setting.