Fetching from the wire…
01
02
03
04
05
06
07
08
Models2026-09-07 · source-backed
A hardware-aware pipeline combining structured pruning, quantization-aware training and fixed-length bit-packing reaches a 55.8x weight compression ratio on LeNet-5 at 98.77% accuracy, emitting a dependency-free C library with deterministic latency (arXiv 2609.05081). On the RP2040 it cuts binary size 3x against TensorFlow Lite at matching accuracy. The headline is the first documented deployment of a standard CNN on a device previously considered infeasible for CNN inference.
Each link below shares sources, entities, or timing with this story.
Speculative decoding usually splits within one machine. SPADE splits across the edge and cloud boundary: a compact draft model on the edge proposes tokens, a large cloud verifier validates them in parallel, and only rejections trigger a cloud correction (arXiv). No retraining,...
Translating key-value state from one model into a form another can consume works across scale, architecture, attention configuration, tokenizer and family (arXiv 2608.30963). Llama3.1-70B to Qwen2.5-7B reaches 44.0% accuracy against 45.7% native while dropping latency to 138ms...
The study extracted 130 clean atomic state transitions from 707 real issues in SWE-bench Lite and Verified. Plain RAG scored 0.57-0.59 answer accuracy; an LLM reranker didn't help and added latency, about 18 seconds against 2.1. A (subject, relation, object) supersession memor...
RGA-Designer trains a reward model scoring both task correctness and structural compactness, then fine-tunes a graph generator against it to design communication topologies. arXiv For fan-out agent teams where inter-agent chatter dominates the bill, topology is a cost lever mo...
arXiv 2607.26935 argues the human-vs-bot label space can't represent agent traffic: an MLP binary classifier misroutes 39.1% of real agent sessions as human, a SAINT transformer 34.5%, while adding an explicit third class yields agent F1 = 1.000 across all 30 runs. Against a f...
D-SCAN (SIGIR 2026) found the standard guardrail returns high confidence on compromised output. Their alternative signal is document-level attention dynamics: during a poisoned generation, attention concentrates on the injected document and entropy collapses, versus dispersed...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.