AI Low Power - Revision history

Tom: /* Ethical and Governance */

2025-10-10T04:37:07Z

‎Ethical and Governance

2025-10-10T04:36:19Z

‎Key Features

2025-10-10T04:35:28Z

‎Technical Design Patterns

2025-09-11T19:22:43Z

‎New Low Power Paradigm

2025-09-11T19:20:07Z

‎Spiking

2025-09-11T19:19:05Z

‎New Low Power Paradigm

2025-08-24T22:32:26Z

‎References

2025-08-24T22:31:57Z

‎Developer Tools

2025-07-10T22:02:32Z

‎New Low Power Paradigm

2025-07-10T21:59:13Z

‎New Low Power Paradign

@@ Line 53: / Line 53: @@
 ===Ethical and Governance===
 Low-power AIs often serve **underserved regions or infrastructure-poor contexts**. You could integrate:
-- **Consent-aware identity presentation** (aligned with OpenID4VP)
+*Consent-aware identity presentation** (aligned with OpenID4VP)
-- **Auditable interactions without surveillance**
+*Auditable interactions without surveillance**
-- **Localized model training** to respect cultural data boundaries
+*Localized model training** to respect cultural data boundaries
 ==Tensor Flow Lite==

@@ Line 61: / Line 61: @@
 ===Key Features===
-- **Optimized for low power and latency**: Ideal for real-time inference on devices with limited compute and memory.
+*Optimized for low power and latency**: Ideal for real-time inference on devices with limited compute and memory.
-- **Offline capability**: No need for cloud access—models run locally.
+*Offline capability**: No need for cloud access—models run locally.
-- **Small binary size**: Uses the `.tflite` format (FlatBuffers) for compact deployment.
+*Small binary size**: Uses the `.tflite` format (FlatBuffers) for compact deployment.
-- **Cross-platform support**: Works on Android, iOS, embedded Linux, and microcontrollers.
+*Cross-platform support**: Works on Android, iOS, embedded Linux, and microcontrollers.
-- **Hardware acceleration**: Supports GPU, NNAPI, and Core ML delegates for faster performance.
+*Hardware acceleration**: Supports GPU, NNAPI, and Core ML delegates for faster performance.
 ===How It Works===

@@ Line 28: / Line 28: @@
 ===Technical Design Patterns===
-- **TinyML Models**: Leverage frameworks like [TensorFlow Lite Micro](https://www.tensorflow.org/lite/microcontrollers) or [Edge Impulse](https://www.edgeimpulse.com/) to deploy on microcontrollers.
+*TinyML Models**: Leverage frameworks like [https://www.tensorflow.org/lite/microcontrollers TensorFlow Lite Micro] or [https://www.edgeimpulse.com/ Edge Impulse] to deploy on micro-controllers.
-- **Quantized Inference**: Use int8 or int4 precision models to drastically cut power and memory consumption.
+*Quantized Inference**: Use int8 or int4 precision models to drastically cut power and memory consumption.
-- **Event-Driven Architecture**: Wake the AI only on sensor triggers (e.g., sound, movement), using interrupt logic.
+*Event-Driven Architecture**: Wake the AI only on sensor triggers (e.g., sound, movement), using interrupt logic.
-- **BLE/NFC Integration**: Avoid constant connectivity; use short-range communication for burst interaction.
+*BLE/NFC Integration**: Avoid constant connectivity; use short-range communication for burst interaction.
-- **Rule-Based Fallbacks**: Combine ML with deterministic logic for systems where full model inference is too costly.
+*Rule-Based Fallbacks**: Combine ML with deterministic logic for systems where full model inference is too costly.
 ===Ethical and Governance===

@@ Line 74: / Line 74: @@
 Explore more in IBM’s research blog.
 ===Spiking Neural Networks (SNNs) from IIT Bombay===
 Researchers developed a chip using band-to-band tunneling (BTBT) to power spiking neural networks. These mimic brain-like behavior and consume 5,000× less energy per spike than conventional designs. Ideal for speech recognition, motion detection, and biomedical signals on edge devices.
-Read the IEEE Spectrum article here.
+Read the IEEE Spectrum article [https://spectrum.ieee.org/low-power-ai-spiking-neural-net here.]
 ===Falcons.AI + Intel Liftoff===
@@ Line 93: / Line 100: @@
 You can read the full research paper [https://link.springer.com/article/10.1007/s11432-023-3813-8 here].
 ==References==
 [[Category: Artificial Intelligence]]
 [[Category: Networking]]

@@ Line 94: / Line 94: @@
 You can read the full research paper [https://link.springer.com/article/10.1007/s11432-023-3813-8 here].
-==Spiking==
+===Spiking===
 SpikingBrain1.0 is an important development, but the headlines need context. Spiking neural networks can be more efficient than transformers, especially on very long sequences where quadratic complexity bogs models down. In those scenarios, speedups of “100×” are possible. But that doesn’t mean this replaces today’s LLMs, for most real-world tasks, the gains are smaller, and the tradeoffs in accuracy and robustness aren’t clear.

← Older revision		Revision as of 22:32, 24 August 2025
Line 96:		Line 96:
	==References==		==References==
	[[Category: Artificial Intelligence]]		[[Category: Artificial Intelligence]]
		+	[[Category: Networking]]

@@ Line 67: / Line 67: @@
 * [LiteRT](https://ai.google.dev/edge/litert): The next-gen runtime evolving from TensorFlow Lite, with broader model support and improved acceleration.
-Thhinking about deploying low-power AI for identity systems or BLE/NFC flows, TensorFlow Lite is a solid foundation. Want help setting up a model or choosing the right delegate for your platform? Thinking of prototyping, scaffold out a repo structure or design specs. Or explore something like **mesh-networked credential validation**
+Thinking about deploying low-power AI for identity systems or BLE/NFC flows, TensorFlow Lite is a solid foundation. Want help setting up a model or choosing the right delegate for your platform? Thinking of prototyping, scaffold out a repo structure or design specs. Or explore something like [[Mesh Network]] credential validation**
 ==New Low Power Paradigm==