#Optimization

SkewAdam: Rethinking MoE Optimizer Memory

SkewAdam drastically cuts MoE training memory by tailoring optimizer state to parameter populations, achieving superior perplexity and enabling training on accessible hardware.

Hardware-Software Co-Design: AI's 100x Multiplier

Dylan Patel of SemiAnalysis explains how hardware-software co-design is the key to unlocking 100x performance gains in AI, optimizing data flow and reducing costs.

Uber's Knapsack Problem

Uber's Tarot platform employs a sophisticated Multiple Knapsack Problem solver to optimize incentive allocation, balancing user experience with strict budget constraints.

Go Stack Allocation Boosts Uber CPU Efficiency

Uber engineers optimized Go stack allocation, cutting CPU usage by up to 10% in critical services through static pre-allocation and disabling dynamic growth.

Microsoft's GridSFM: AI for the Power Grid

Microsoft's new GridSFM AI model drastically speeds up power grid analysis, promising efficiency gains and cost savings.

Parallelizing Sequential Dynamics

New parallel Newton methods overcome sequential bottlenecks in dynamical systems, offering stability and provable acceleration guarantees.

Agentic LLMs: Stabilizing Minimax Training

Adversarially-Aligned Jacobian Regularization (AAJR) tackles LLM agent stability by controlling sensitivity along adversarial directions, expanding policy classes and reducing performance degradation.

Small language model optimization cracks complex business math

Microsoft’s OptiMind is a 20-billion parameter small language model that achieves high accuracy in converting natural language business problems into mathematical optimization models through expert-aligned training.

AI Wins Optimization Contest, Costs $1,300 to Beat Humans

UCSD Lab Advances Low-Latency LLM Serving with DGX B200

AI startup Rookoo funding targets event admin chaos

The Unseen Challenge of Reliable AI

Fal.ai's Blueprint for AI Video Dominance: Speed, Specialization, and Relentless Optimization

Tibo Energy Secures $6.5 Million in Seed Funding

Tibo Energy, an Eindhoven-based developer of AI-driven software for industrial and commercial energy management, has secured $6.5 million in Seed funding.

Tibo Energy Secures $6.5 Million in Seed Funding

Tibo Energy, an Eindhoven-based developer of AI-driven software for industrial and commercial energy management, has secured $6.5 million in Seed funding.

Sedai Secures $20 Million Series B to Revolutionize Cloud Optimization with AI Agents

The company's platform currently manages over $1.2 billion in cloud spend for its diverse customer base, which includes industry giants like Palo Alto Networks and HP.

Speedata Secures $44M Series B Funding to Revolutionize Data Analytics

Future AGI Secured $1.6M to Optimize Agentic AI Workflows

As enterprise AI adoption accelerates, a staggering 85% of AI projects fail to meet expectations due to accuracy and reliability challenges.

Future AGI Secured $1.6M to Optimize Agentic AI Workflows

As enterprise AI adoption accelerates, a staggering 85% of AI projects fail to meet expectations due to accuracy and reliability challenges.

enSights Secures $10M in Series A Funding for Clean Energy Optimization Expansion

3D Printing Startup CASTOR Raises Funding From Asahi Kasei

Accelerating Deep Learning: Transforming Batch Processing into Real-Time Mastery

How to Reduce Multi-Cloud Costs with Kyndryl and CloudWize

Making Machine Learning Inference Meet Real-World Performance Demands

FPGAs offer the configurability needed for real-time machine learning inference, with the flexibility to adapt to future workloads. Making these advantages accessible to data-scientists and developers calls for tools that are both comprehensive and easy to use. Daniel Eaton, Sr Manager, Strategic Marketing Development, Xilinx