đź§ Daily Study Log [2025-10-07]
Focused on coursework across AI and data-related subjects, proposed a new post-holiday recovery idea, and began studying InternVideo2, a recent multimodal video foundation model.
đź“„ Coursework / Activities
- Deep Learning
- Transitioned from manually built binary classifiers to model design using deep learning libraries.
- Blockchain & Cryptocurrency
- Explored evolution from Proof-of-Work to private and hybrid coins, deepening understanding of modern crypto mechanisms.
- Big Data Analysis Practice
- Studied statistical fundamentals critical for data analysis, including chi-square and Poisson distributions.
- Understanding Cryptography
- Compared classical vs. modern encryption, identifying principles behind current cryptographic standards.
- Positive Psychology
- Learned how ownership vs. experience affects happiness and practiced techniques like mindfulness, walking meditation, and small joys.
- AI-based Havruta Education
- Examined five major Havruta (dialogue-based learning) models and their educational applications.
- AI Creative Convergence Strategy
- Studied generative AI and its role in emerging creative industries and digital content creation.
đź’ˇ Idea
- Proposed Social Recovery After Holiday.
- Concept: Encourage post-holiday emotional reset and community motivation through shared recovery activities.
- Goal: Design small, collective habits that ease transition back to daily life after long holidays.
📖 Paper Study – InternVideo2 (CVPR 2024)
- Reviewed InternVideo2: Scaling Video Foundation Models for Multimodal Understanding.
- Key insights:
- Introduces a unified framework combining masked video modeling, contrastive learning, and next-token prediction.
- Focuses on spatiotemporal coherence and cross-modal (video–text–audio–speech) alignment.
- Trained on 400M video-text pairs, achieving SOTA in video retrieval, captioning, and action recognition.
- Takeaway: InternVideo2 demonstrates how scalable multimodal architectures can generalize across diverse video-language tasks.
âś… TL;DR
📍 Completed multiple AI-related coursework sessions
📍 Proposed “Social Recovery After Holiday” idea for emotional reset and motivation
📍 Studied InternVideo2 — unified, scalable, multimodal video foundation model