Skip to content

Top Stories

Top Stories

Primary Menu
  • Breaking News
  • UNIT CONVERTER
  • QR Code Generator
  • SEO META TAG GENERATOR
  • Background Remover Tool
  • Image Enhancer Tool
  • Image Converter Tool
  • Image Compressor Tool
  • Keyword Research Tool
  • Paint Tool
  • About Us
  • Contact Us
  • Privacy Policy
HOME PAGE
  • Home
  • Uncategorized
  • This AI Paper Introduces Inference-Time Scaling Techniques: Microsoft’s Deep Evaluation of Reasoning Models on Complex Tasks
  • Uncategorized

This AI Paper Introduces Inference-Time Scaling Techniques: Microsoft’s Deep Evaluation of Reasoning Models on Complex Tasks

VedVision HeadLines April 8, 2025
This AI Paper Introduces Inference-Time Scaling Techniques: Microsoft’s Deep Evaluation of Reasoning Models on Complex Tasks


Large language models are often praised for their linguistic fluency, but a growing area of focus is enhancing their reasoning ability—especially in contexts where complex problem-solving is required. These include mathematical equations and tasks involving spatial logic, pathfinding, and structured planning. In such domains, models must simulate human-like step-by-step thinking, where solutions are not immediately obvious. This type of structured reasoning makes inference-time behavior an important subject of study in machine learning research.

Despite the progress in model architecture and training datasets, many language models still falter when presented with multi-step or high-difficulty reasoning tasks. The challenge is that even if a model can access vast information, it might not know how to use it effectively across multiple steps. Tasks like selecting meeting times with constraints or solving NP-hard problems require sustained logical sequencing, which standard models find difficult. Adding more parameters or memory has helped in some areas, but such brute-force solutions often lead to diminishing returns when task complexity increases.

To handle these limitations, researchers have explored tools like chain-of-thought prompting and post-training fine-tuning to better align models with complex tasks. Some methods involve generating multiple independent answers and then using heuristics or voting mechanisms to pick the most likely correct one. Others experiment with self-refinement—having the model critique its answers and revise accordingly. These approaches have been implemented with varying success in conventional models such as GPT-4o, Claude 3.5 Sonnet, and Gemini 2.0 Pro, but these models still show variability depending on the benchmark. In some instances, longer output did not translate into better accuracy, and token efficiency remained inconsistent.

Researchers at Microsoft introduced a rigorous evaluation framework for inference-time scaling that covers nine models and eight complex task benchmarks. This included comparing conventional models against reasoning-optimized ones such as DeepSeek R1, O1, and O3-mini. Their method involved parallel scaling, where multiple outputs are generated and aggregated, and sequential scaling, where the model is prompted to revise its output based on structured feedback iteratively. Benchmarks were sourced from domains like calendar planning, math Olympiads, and spatial reasoning, and the team introduced two new datasets for NP-hard problems: 3SAT and TSP.

The methodology relied on two core strategies: sampling multiple generations to evaluate result variability and using critics to simulate feedback-enhanced reasoning. In parallel scaling, the model outputs several answers that are evaluated using aggregators such as majority vote or best-of-n. In sequential scaling, the model receives feedback after each attempt and is prompted to try again. This allowed researchers to estimate current performance and the potential ceiling for improvement if computational resources were scaled up. Aggregators like average and worst-of-n helped identify where models consistently failed or succeeded. This dual approach provided insight into how models use additional inference steps and whether feedback mechanisms improve answer quality.

The performance analysis showed significant differences between models and task types. On the GPQA benchmark, the top-performing model, O1, reached 90.9% accuracy, while GPT-4o reached 77.7%. On the TSP dataset, O1 maintained accuracy above 80% across most levels, while GPT-4o’s performance peaked only when superscaled with over 20 inference calls. In BA Calendar, DeepSeek R1 achieved 88.5% accuracy, outperforming Claude 3.7 Sonnet and Gemini 2.0 Pro. However, results also revealed that increased token usage did not guarantee higher accuracy. For example, DeepSeek R1 consumed significantly more tokens than Claude 3.7 Sonnet but only marginally outperformed it in some math tasks. Even within a single model, repeated attempts on the same question showed high variation in token counts, raising concerns about cost predictability for real-world applications.

This study underscores the gap between traditional and reasoning-enhanced models and highlights that intelligent scaling—not just more tokens—can improve complex task performance. The researchers showed that feedback loops and strong verifiers offer substantial gains in model accuracy, even in difficult domains. Their findings suggest that reasoning models still have headroom for improvement, especially when guided by structured inference strategies and cost-efficient token management.


Check out the Paper and GitHub. All credit for this research goes to the researchers of this project. Also, feel free to follow us on Twitter and don’t forget to join our 85k+ ML SubReddit.

🔥 [Register Now] miniCON Virtual Conference on OPEN SOURCE AI: FREE REGISTRATION + Certificate of Attendance + 3 Hour Short Event (April 12, 9 am- 12 pm PST) + Hands on Workshop [Sponsored]


Nikhil is an intern consultant at Marktechpost. He is pursuing an integrated dual degree in Materials at the Indian Institute of Technology, Kharagpur. Nikhil is an AI/ML enthusiast who is always researching applications in fields like biomaterials and biomedical science. With a strong background in Material Science, he is exploring new advancements and creating opportunities to contribute.



Source link

Continue Reading

Previous: Waqf Bill Protest Update; BJP NC Congress | Jammu Kashmir Muslims – Supreme Court | जम्मू-कश्मीर विधानसभा में नेशनल कॉन्फ्रेंस और भाजपा विधायकों में हाथापाई: नए वक्फ कानून को लेकर भिड़े; कल बिल की कॉपी फाड़ी थी
Next: China’s tariff response may mean more capital flight to crypto: Hayes

Related News

Jefferies Raises Disney Stock to Buy With 4 Price Target
  • Uncategorized

Jefferies Raises Disney Stock to Buy With $144 Price Target

VedVision HeadLines July 6, 2025
Will SOL rally to 0?
  • Uncategorized

Will SOL rally to $200?

VedVision HeadLines July 6, 2025
AbstRaL: Teaching LLMs Abstract Reasoning via Reinforcement to Boost Robustness on GSM Benchmarks
  • Uncategorized

AbstRaL: Teaching LLMs Abstract Reasoning via Reinforcement to Boost Robustness on GSM Benchmarks

VedVision HeadLines July 6, 2025

Recent Posts

  • 12-year-old girl gang-raped, murdered in Uttar Pradesh village, four booked
  • Jefferies Raises Disney Stock to Buy With $144 Price Target
  • Will SOL rally to $200?
  • How a Canadian’s AI hoax duped the media and propelled a ‘band’ to streaming success
  • BBC Antiques Roadshow expert admits ‘I’ve never seen one’ as he declares item ‘one of the rarest ever’

Recent Comments

No comments to show.

Archives

  • July 2025
  • June 2025
  • May 2025
  • April 2025

Categories

  • Current Affairs
  • Shopping
  • Uncategorized

You may have missed

12-year-old girl gang-raped, murdered in Uttar Pradesh village, four booked
  • Current Affairs

12-year-old girl gang-raped, murdered in Uttar Pradesh village, four booked

VedVision HeadLines July 6, 2025
Jefferies Raises Disney Stock to Buy With 4 Price Target
  • Uncategorized

Jefferies Raises Disney Stock to Buy With $144 Price Target

VedVision HeadLines July 6, 2025
Will SOL rally to 0?
  • Uncategorized

Will SOL rally to $200?

VedVision HeadLines July 6, 2025
How a Canadian’s AI hoax duped the media and propelled a ‘band’ to streaming success
  • Current Affairs

How a Canadian’s AI hoax duped the media and propelled a ‘band’ to streaming success

VedVision HeadLines July 6, 2025
Copyright © All rights reserved. | MoreNews by AF themes.