# Cognitive Services

Published articles for Cognitive Services.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Exploring Microsoft's AI foundation models

DevFeed: [Exploring Microsoft's AI foundation models](<https://devfeed.tech/articles/exploring-microsoft-s-ai-foundation-models-130248.md>)

Original publisher: [Read original article](<https://www.infoworld.com/article/4245368/exploring-microsoft-ai-foundation-models.html>)

Author: Simon Bisson

Published: 2026-11-12T09:00:00Z

Content type: article

Language: en

Sources: [InfoWorld](<https://devfeed.tech/sources/infoworld.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Azure OpenAI](<https://devfeed.tech/topics/azure-openai.md>), [Frontier AI](<https://devfeed.tech/topics/frontier-ai.md>), [Azure](<https://devfeed.tech/topics/azure.md>), [OpenAI](<https://devfeed.tech/topics/openai.md>), [Claude](<https://devfeed.tech/topics/claude.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-data-centers](<https://devfeed.tech/tags/ai-data-centers.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [artificial-intelligence-cloud-computing-generative-ai-iaas-microsoft-azure](<https://devfeed.tech/tags/artificial-intelligence-cloud-computing-generative-ai-iaas-microsoft-azure.md>), [azure](<https://devfeed.tech/tags/azure.md>), [azure-openai](<https://devfeed.tech/tags/azure-openai.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cloud-computing](<https://devfeed.tech/tags/cloud-computing.md>), [cognitive-services](<https://devfeed.tech/tags/cognitive-services.md>), [computer-vision](<https://devfeed.tech/tags/computer-vision.md>), [data-centers](<https://devfeed.tech/tags/data-centers.md>), [deep-learning](<https://devfeed.tech/tags/deep-learning.md>), [deepseek](<https://devfeed.tech/tags/deepseek.md>), [edge](<https://devfeed.tech/tags/edge.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [foundry](<https://devfeed.tech/tags/foundry.md>), [frontier-models](<https://devfeed.tech/tags/frontier-models.md>), [generative-ai](<https://devfeed.tech/tags/generative-ai.md>), [google](<https://devfeed.tech/tags/google.md>), [iaas](<https://devfeed.tech/tags/iaas.md>), [language-en](<https://devfeed.tech/tags/language-en.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>), [microsoft-azure](<https://devfeed.tech/tags/microsoft-azure.md>), [models](<https://devfeed.tech/tags/models.md>), [storytype-analysis](<https://devfeed.tech/tags/storytype-analysis.md>)

### AI overview

Microsoft's MAI program adds its own foundation models to Azure Foundry, emphasizing lower inference cost and power use while aiming for results that are sufficient for common tasks. The models span cybersecurity, coding, transcription, and a Thinking model; MAI also offers tuning on customer data. MAI-Thinking-1 can be deployed through existing Foundry Azure OpenAI APIs, with token limits, tool calls, encrypted reasoning chains, and streaming. Flash models target local or cloud use: MAI-Cyber-1.1-Flash reportedly handles 90% of MDASH tasks previously assigned to larger models and costs 18% less to run than version 1.0, escalating harder work to a more powerful model. MAI-Code-1.1-Flash requires over 120 GB of RAM, limiting local use to high-memory hardware. Microsoft's Maia 300 inference platform is designed for its models and OpenAI; the article reports 36% more compute, 2.5x higher memory bandwidth, and roughly 2x performance per watt versus the prior generation. The article frames lower operating cost, rather than matching the largest models in every respect, as Microsoft's competitive case.

### Source excerpt

When Microsoft started to deploy AI models in what is now its Foundry platform, it was building on its partnership with OpenAI. OpenAI was able to use Azure infrastructure to train its models, and in return, Microsoft became the first commercial partner to use them. Azure OpenAI has been a successful product for both businesses. But as more organizations got access to OpenAI models, Microsoft's stable of models grew to include not only OpenAI's, but also models from Anthropic and Hugging Face, as well as models from Meta, Mistral, and DeepSeek. At the same time, Microsoft has continued to offer its own Cognitive Services models for computer vision and text processing, as well as the Phi series of efficient lightweight models for use on the edge. With access to more than 11,000 different models, Foundry has proven to be a useful way to build modern AI-based applications, offering many agent-focused features, including tools to ground models using Microsoft's graph-based IQ data sources. But one key element has been missing -- Microsoft's own large-scale frontier models. Microsoft AI has a long history It's not that Microsoft is new to developing its own AI models. Microsoft Research has been a hub for deep learning for much of the last decade, with models that regularly top benchmark tables. But it hasn't taken the same general-purpose approach as Anthropic, OpenAI, or even Google, instead producing more focused models. Now Microsoft has begun to roll out its own foundation models as part of its Microsoft AI (MAI) program. With a focus on offering lower token cost than competitors, MAI has spotted a gap in the market. It's a sensible move to make, too. The days of free and low-cost tokens have gone, never to return, as hyperscale vendors look to recoup the massive multi-billion-dollar investments they've made in AI data centers. So-called "tokenmaxxing" is no longer economically viable, dumping massive data sets into LLMs with the hope of getting results. Instead, inf

## Building a Multimodal LLM Pipeline for Structured Visual-Asset Metadata

DevFeed: [Building a Multimodal LLM Pipeline for Structured Visual-Asset Metadata](<https://devfeed.tech/articles/teaching-a-vision-model-to-see-like-a-human-annotator-and-catching-it-when-it-lies-50201.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/ise/ai-asset-enrichment-pipeline/>)

Author: Deeptanil Saha, Kartheek Palepu

Published: 2026-07-17T07:00:00Z

Content type: article

Language: en

Sources: [ISE Developer Blog](<https://devfeed.tech/sources/ise-developer-blog.md>)

Topics: [multimodal-ai](<https://devfeed.tech/topics/multimodal-ai.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Ground truth / benchmark quality](<https://devfeed.tech/topics/ground-truth-benchmark-quality.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>)

Tags: [cognitive-services](<https://devfeed.tech/tags/cognitive-services.md>), [cse](<https://devfeed.tech/tags/cse.md>), [developer](<https://devfeed.tech/tags/developer.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [hallucinations](<https://devfeed.tech/tags/hallucinations.md>), [ise](<https://devfeed.tech/tags/ise.md>), [llm](<https://devfeed.tech/tags/llm.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [model](<https://devfeed.tech/tags/model.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [structured](<https://devfeed.tech/tags/structured.md>), [vision](<https://devfeed.tech/tags/vision.md>)

### AI overview

This article describes a multimodal LLM pipeline that extracts structured metadata from visual assets for production search. It uses constrained values, image-first grounding, preprocessing, and evaluation to reduce hallucinations and improve trust in the resulting metadata.

### Source excerpt

A multimodal LLM enrichment pipeline that extracts structured metadata from visual assets, constrains output to predefined values to minimize hallucinations, and uses a ground truth evaluation template to measure quality--all as a plug-and-play module. The post Teaching a Vision Model to See Like a Human Annotator--and Catching It When It Lies appeared first on ISE Developer Blog.

## From Noisy Queries to Precise Frames: Query Decomposition for Media Asset Search

DevFeed: [From Noisy Queries to Precise Frames: Query Decomposition for Media Asset Search](<https://devfeed.tech/articles/from-noisy-queries-to-precise-frames-query-decomposition-for-media-asset-search-50204.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/ise/from_noisy_queries_to_precise_frames/>)

Author: Kartheek Palepu, Deeptanil Saha

Published: 2026-07-09T07:00:00Z

Content type: article

Language: en

Sources: [ISE Developer Blog](<https://devfeed.tech/sources/ise-developer-blog.md>)

Topics: [Query (disambiguation)](<https://devfeed.tech/topics/query.md>), [Parser](<https://devfeed.tech/topics/parser.md>), [Parsing](<https://devfeed.tech/topics/parsing.md>), [data](<https://devfeed.tech/topics/data.md>), [processing](<https://devfeed.tech/topics/processing.md>)

Tags: [cognitive-services](<https://devfeed.tech/tags/cognitive-services.md>), [cse](<https://devfeed.tech/tags/cse.md>), [data](<https://devfeed.tech/tags/data.md>), [developer](<https://devfeed.tech/tags/developer.md>), [frameworks](<https://devfeed.tech/tags/frameworks.md>), [hybrid](<https://devfeed.tech/tags/hybrid.md>), [ise](<https://devfeed.tech/tags/ise.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [queries](<https://devfeed.tech/tags/queries.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [regex](<https://devfeed.tech/tags/regex.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [search](<https://devfeed.tech/tags/search.md>), [useful](<https://devfeed.tech/tags/useful.md>), [vector](<https://devfeed.tech/tags/vector.md>)

### AI overview

This article explains a query-decomposition approach for media asset search. It separates structured metadata such as episode, scene, and shot identifiers from natural-language visual intent, applies the metadata as filters, and uses semantic or vector retrieval to rank the filtered candidates.

### Source excerpt

How query decomposition separates metadata filters from visual intent to significantly improve media asset retrieval quality. The post From Noisy Queries to Precise Frames: Query Decomposition for Media Asset Search appeared first on ISE Developer Blog.