Contrastive learning has become the dominant paradigm for aligning heterogeneous modalities in multimodal large language models (MLLMs). Information Noise Contrastive Estimation (InfoNCE) Loss and its variants serve as the foundational training objectives for such models. As models scale to billions of parameters and training batches reach millions of samples, the limitations of conventional softmax-based contrastive losses have become increasingly apparent: quadratic memory growth, batch size sensitivity, semantic noise in negative sampling, and the inherent tension between alignment and uniformity. This paper presents a comprehensive survey of the evolution from InfoNCE to next-generation objectives for multimodal LLMs. We analyze the theoretical foundations of contrastive learning through the alignment-uniformity framework, examine the practical innovations that have enabled unprecedented batch scaling, and evaluate emerging alternatives including sigmoid-based losses, non-contrastive alignment methods, and hybrid generative-discriminative objectives.

Large language models (LLMs) have achieved remarkable performance across diverse tasks, yet they remain fundamentally static: once trained, their knowledge is frozen, unable to adapt to evolving information, emerging domains, or shifting user needs without expensive retraining. This limitation - catastrophic forgetting - has become a critical bottleneck as LLMs transition from research artifacts to living systems deployed in dynamic real-world environments. This paper presents a comprehensive survey and unified framework for continuously learning LLMs (CL-LLMs), organizing the field into five complementary paradigms: replay-based methods, regularization-based approaches, parameter-efficient adaptation, architectural isolation, and inference-time continual learning.

The proliferation of sovereign large language models (LLMs) trained on indigenous languages, cultural contexts, and national datasets has created an urgent need for inference hardware that preserves the unique characteristics of these models while achieving deployment efficiency. Unlike generic AI accelerators optimized for English-centric transformer models, sovereign inference chips must handle multilingual tokenization, mixed-script processing, sparse mixture-of-experts (MoE) architectures, and voice-first interaction patterns that dominate local language AI deployments. This paper presents a model-driven co-design framework that maps the specific computational signatures of sovereign AI models directly to custom inference chip architectures.

The IndiaAI–NCG CATCH Grant for Cancer 2026 compendium documents ten winning AI solutions selected through a competitive grant program jointly run by IndiaAI (MeitY) and the National Cancer Grid's Koita Centre for Digital Oncology. It profiles AI tools spanning cancer screening, diagnostics, clinical decision support, patient engagement, and operational efficiency, ranging from pilot-ready to commercially deployed systems. Each solution is evaluated against criteria including clinical relevance, technical maturity, data governance, DPDP compliance, and deployment feasibility within Indian hospital settings. The compendium also covers the selection methodology, jury composition, responsible AI principles, and key learnings from the challenge, serving as a reference for hospitals, policymakers, and innovators seeking to adopt or scale AI-enabled cancer care across India.

This article highlights how AI-powered image validation in PCSAP is improving India’s grain procurement monitoring system. With thousands of procurement centers and nearly 14 lakh photographs uploaded every season, manual verification had become difficult and time-consuming. The AI system validates images in real time by checking location, timestamp, quality, duplication, and content relevance. This helps improve transparency, reduce manual effort, enhance data accuracy, and support faster decision-making in procurement center assessments. The solution has been piloted in Punjab, Haryana, and Odisha, with potential for wider use in government monitoring and compliance programs.

This paper proposes a comprehensive framework for integrating LoRA adapters into sovereign AI repositories such as AIKosh. Instead of treating each domain-specific model as an independent artifact, the proposed architecture stores foundation models separately from lightweight LoRA adapters, enabling modular deployment, efficient storage, rapid domain adaptation, and improved model governance. The paper introduces a repository architecture supporting adapter versioning, standardized metadata, benchmarking, interoperability, GPU-enabled training workflows, and community-driven evaluation. It discusses the role of LoRA repositories in strengthening AI sovereignty by enabling country-specific AI capabilities to be developed, shared, and continuously improved without duplicating large foundation models. The proposed framework has the potential to reduce computational costs, democratize AI development for startups and academic institutions, encourage collaborative innovation, and establish AIKosh as a national ecosystem for modular AI capability development.

Deepfake detection looks solved on paper. 0.98 AUC on FaceForensics++. Then 0.65 on Celeb-DF. That gap is the real story. This article breaks down why passive detection fails across datasets, why diffusion-generated fakes are structurally harder to catch than GAN-based ones, and what India's new IT Rules 2026 demand from platforms technically. It also maps the open problems: cross-dataset generalisation, compression robustness, and the complete absence of Indian-face deepfake benchmarks. Written for ML practitioners, not headlines.

The article explores how China is aggressively advancing its geopolitical influence by releasing powerful, open-source AI models, making high-tier AI capabilities freely and globally accessible. Contrasts this with the proprietary, closed-source approaches dominated by US tech giants. The piece analyzes India's strategic position within this shifting, multipolar AI ecosystem, arguing that India must carefully balance its infrastructure dependencies, leverage open-source breakthroughs for localized economic growth, and craft its own independent AI policies to avoid being caught in a digital duopoly between the US and China.

Why owning both the silicon and the software is the only durable moat in the age of AI, and why the margin arithmetic of fragmented supply chains is finally breaking down. By Dhritiman Mallick · Founder & CEO, Vyuhaa Med Data

AI coding assistants are impressive on greenfield projects, but most real-world engineering happens inside brownfield systems: legacy codebases with weak tests, hidden dependencies, and accumulated complexity. This article introduces “Agentic Code Surgery,” a seven-agent workflow that helps AI assistants change such systems safely by first understanding behavior, identifying seams, creating characterization tests, and only then implementing changes. The paper is attached for readers who want the full method, experiment, and evidence.

Dataset distillation is an emerging technique that compresses large-scale training datasets into small, informative synthetic or selected subsets while preserving the essential knowledge required for training high-performing machine learning models. This paper analyses the techniques in dataset distillation, with a particular focus on selection-based methods, and examines how these techniques can address AIKosh’s core challenges related to efficient storage, reduced computational costs, cross-architecture generalization, and democratized access to high-quality training data. The research proposes that integrating dataset distillation into AIKosh’s ecosystem can help accelerate India’s sovereign AI mission by enabling resource-constrained researchers and startups to train competitive models on compressed, representative data.

Northeast India is home to over 220 languages spoken by 45 million people, yet remains mostly absent from modern AI systems. MWire Labs, a northeast India AI startup based in Shillong, Meghalaya, is building the NE-Stack, a foundational suite of northeast AI models covering speech recognition, machine translation, OCR, TTS, and vision-language understanding across 11+ indigenous languages. This article documents the technical and strategic approach behind northeast AI development for one of the world's most linguistically diverse and digitally underserved regions.