Ravidu Silva

Out of the numerous types of Medical Imaging modalities available, radiography stands out a bit more than others due to its capabilities of diagnosing diseases and conditions, including life-threatening conditions. Its affordability is another main reason for its prevalence. Chest Radiography holds even higher importance, as it focuses a critical area of the human body. However, interpreting a Chest Radiography image can be challenging and usually done by an experienced Radiologist for accurate results. There are two main issues related to this. One is that in some countries, experienced Radiologists are scarce. The other issue is that the inevitability of human errors in diagnoses. Researchers attempt to use Artificial Intelligence to address these two issues. Most of the existing work incorporates Convolutional Neural Networks for this purpose. This paper presents a novel way of parallelizing multiple architectures of Convolutional Neural Networks focusing on Chest X-ray classification. The paper further presents a comprehensive evaluation of the existing architectures with the parallelized results of them using our method. We used four large-scale datasets, including a non-medical one, for the evaluation of our models. We managed to achieve better accuracy for 9 out 13 and 11 out of 14 labels on our two main evaluation datasets. The paper concludes by presenting the limitations and future improvements possible for the system.

Generating programmatic animation using libraries such as Manim presents unique challenges for Large Language Models (LLMs), requiring spatial reasoning, temporal sequencing, and familiarity with domain-specific APIs that are underrepresented in general pre-training data. A systematic study of how training and inference strategies interact in this setting is lacking in current research. This study introduces ManimTrainer, a training pipeline that combines Supervised Fine-tuning (SFT) with Reinforcement Learning (RL) based Group Relative Policy Optimisation (GRPO) using a unified reward signal that fuses code and visual assessment signals, and ManimAgent, an inference pipeline featuring Renderer-in-the-loop (RITL) and API documentation-augmented RITL (RITL-DOC) strategies. Using these techniques, this study presents the first unified training and inference study for text-to-code-to-video transformation with Manim. It evaluates 17 open-source sub-30B LLMs across nine combinations of training and inference strategies using ManimBench. Results show that SFT generally improves code quality, while GRPO enhances visual outputs and increases the models’ responsiveness to extrinsic signals during self-correction at inference time. The Qwen 3 Coder 30B model with GRPO and RITL-DOC achieved the highest overall performance, with a 94% Render Success Rate (RSR) and 85.7% Visual Similarity (VS) to reference videos, surpassing the baseline GPT-4.1 model by +3 percentage points in VS. Additionally, the analysis shows that the correlation between code and visual metrics strengthens with SFT and GRPO but weakens with inference-time enhancements, highlighting the complementary roles of training and agentic inference strategies in Manim animation generation.

Large Language Model (LLM) based AI educational content generation systems are increasingly being developed, yet no standardised benchmark exists to systematically evaluate them. This study introduces LessonBench-V1, a benchmark dataset comprising 647 human-written lessons paired with LLM-based reverse-engineered lesson plans across 240 STEM topics spanning mathematics, physics, chemistry, and computer science. The lessons are drawn from 97 trusted open sources, including LibreTexts, Brilliant.org and GeeksForGeeks. Each lesson plan is human-reviewed and produced through a pedagogically grounded methodology that synthesises Bloom’s Taxonomy, Gagné’s Events, Merrill’s First Principles, and the 5E Instructional Model. The lesson plans capture 3,620 learning objectives with pedagogical metadata, enabling systematic, reproducible evaluation of lesson-generation AI agents and supporting further research. The study further proposes a three-dimensional evaluation pipeline for use with the dataset.

The automated generation of educational videos based on Large Language Models (LLMs) holds promising potential to enhance pedagogy. This study investigates the use of compact LLMs with fewer than 10 billion parameters, utilising Parameter-Efficient Fine-Tuning (PEFT) techniques to generate Manim Python code that produces images and videos for educational visualisations from natural language narratives. This approach is highly resource-efficient compared to diffusion-based video generation, which is often infeasible on consumer-level hardware. Due to the lack of data availability, this study curated and reviewed the first benchmarking dataset for Manim code generation, ManimBench. ManimBench features 417 paired descriptions and corresponding Manim code samples categorised as basic, intermediate, and advanced, sourced directly from the official ManimCE documentation. The curated dataset was then used to fine-tune seven open-source compact LLMs using LoRA and QLoRA. Subsequently, their performances were evaluated comparatively against their base models using both functional and semantic metrics. The results indicate that the LoRA Fine-Tuned (FT) models consistently outperform the base models, achieving a 62% Manim render success rate. By leveraging ManimBench, further comparisons were made between the FT models and larger 14B parameter LLMs, finding that the PEFT-tuned FT coder models perform comparably well in accurate Manim code generation. However, it was also discovered that LLMs frequently make hallucination-related API errors in the generated code.

Generating programmatic animations with libraries such as Manim poses unique challenges for large language models (LLMs), requiring spatial reasoning, temporal sequencing, and adaptation to interactive execution environments. However, how different post-training objectives influence an LLM’s capacity to leverage dynamic feedback during generation remains under-researched. This study introduces ManimTrainer, a framework that combines Supervised Fine-tuning (SFT) with visually grounded Reinforcement Learning via Group Relative Policy Optimisation (GRPO) using a unified code-and-visual reward signal, alongside ManimAgent, an inference pipeline featuring Renderer-in-the-loop (RITL) and documentation-augmented execution feedback. Evaluating 17 open-source sub-30B LLMs across nine combinations of training and inference strategies on ManimBench, this work presents a unified study investigating how training types interact with test-time execution feedback. The results show that SFT and GRPO improve visual quality to a comparable degree under vanilla inference, while GRPO also preserves the policy’s responsiveness to extrinsic error signals during self-correction. Under renderer feedback, GRPO outperforms the base and SFT models by median margins of 2.7 pp and 1.3 pp across the 17 models, and the same trend also holds under RITL-DOC. Additionally, the analysis shows that the correlation between code and visual metrics strengthens with SFT and GRPO but weakens under inference-time execution loops, highlighting the complementary roles of training objectives and interactive environments in programmatic generation.