Edge-Ai

TL;DR for operators FedPix1 changes what mobile devices contribute to federated image classification. Rather than training local models, clients perform a single encoding pass and upload compact image-format representations with their labels. The server trains the classifier and updates the shared encoder without receiving the original images. On CIFAR-10, measured client runtime falls from 1,223 seconds with FedAvg to 49 seconds with FedPix, while classification accuracy remains comparable at 65.98% and 65.63%, respectively. The architecture is attractive when local computation is expensive. ...

TL;DR for operators Combining visible and thermal images does not automatically produce a better observation. When corresponding objects appear at different positions, fusion can introduce duplicated edges and structural blur that interfere with detection. SAGE addresses this problem by retaining source-specific information while coordinating appearance equalization, geometric alignment, and image fusion. On the DroneVehicle benchmark, its fused images support vehicle-detection [email protected] of 16.04 versus 12.00 for the strongest competing method, with improvements in both precision and recall. ...

TL;DR for operators A UAV moving repeatedly between suburban and densely built areas encounters communication conditions it has experienced before. Yet a conventional online predictor may have forgotten those conditions while adapting to more recent measurements. LiLib retains small, environment-specific regression models and reactivates them when familiar conditions return. In simulations, this reduces prediction RMSE during the first 50 observations after a revisit from 12.29 to 5.70 dB relative to a sliding-window baseline. The improvement also carries into simulated throughput and outage, while four learned models can be transferred between UAVs using 480 bytes. The operational opportunity is selective model reuse rather than repeated retraining, provided real-world channels exhibit sufficiently recognizable recurring regimes. ...

TL;DR for operators Persistent audiovisual assistants cannot keep extending their working history indefinitely on consumer hardware. As earlier inputs accumulate, the model stores reusable representations so later updates do not have to recompute the entire history; those stored representations consume memory and also increase the attention work required on every update. In one reported Apple M2 Pro profile, MiniCPM-o-4.5 approaches 80K cached tokens and about 23 GiB after 6.25 minutes, while per-chunk prefill latency rises from 28 to 69 seconds and attention time rises from roughly 2 to 39 seconds. ...

TL;DR for operators Fracast-01 reduces a probabilistic time-series foundation model to 85,001 parameters. Against TinyCast, that is a 42.0% parameter reduction, while TinyCast retains 4.2% lower aggregate MASE and 3.3% lower aggregate WQL. The smaller model is therefore an efficiency point, not an accuracy winner. The trade-off is uneven. Fracast-0 is close to TinyCast at short horizons, especially on probabilistic forecast quality, but the gap widens at medium and long horizons. Its weakest reported frequency group is secondly data, where normalized MASE rises above the Seasonal Naive baseline. ...

TL;DR for operators A team choosing between two tiny models for an edge product could reasonably treat the same nominal parameter budget as a fair baseline. In this controlled study at roughly 60K parameters, however, five approximately parameter-matched full-precision transformer shapes span validation losses from 2.835 to 3.477 nats per byte at the 16M-byte training budget—a 22.6% difference produced by the tested shape package alone. Against the weakest transformer shape, the model that dynamically distributes computation across several internal pathways looks roughly 19–21% better; against the strongest full-precision transformer, it is only 0.95% better, inside the study’s seed-noise threshold. ...

TL;DR for operators Speculative decoding saves time when several candidate tokens can be checked together, but the verification step can become its own cost center: if every speculative cycle still traverses the full model, acceleration is bounded by how often that expensive check must run. ECHO1 changes where that verification work happens. It lets cheaper early Transformer layers screen and extend candidates repeatedly, then invokes the remaining layers less frequently for authoritative verification. Intermediate states are reused rather than recomputed. In the paper’s primary comparison table, ECHO reports overall speedups from 2.42× to 2.90× across five model configurations, with overall mean accepted tokens (MAT) from 4.16 to 5.55. ...

TL;DR for operators Lowering an edge processor’s clock can reduce power, but it also slows token generation. That leaves a local LLM runtime with a constrained control problem: keep responses fast enough, stay below thermal limits, limit energy use, and avoid degrading output quality more than the application can tolerate. PELM1 expands the control surface instead of optimizing processor frequency alone. Its runtime governor adjusts CPU and GPU frequencies together with how the model drafts and verifies tokens. On the tested LLaMA-13B workload on Jetson AGX Orin, the paper reports 29.0% to 52.4% lower energy than the evaluated comparison methods across cooling conditions, while PELM delivered the highest average generation speed in three of four conditions. ...

TL;DR for operators A product team normally has several places to stop a dangerous request: filter the prompt, ask another model to inspect it, regenerate under stricter instructions, or moderate the output after generation. All of those controls act around the language model. GUARD-SLM asks whether the model’s own internal representation can provide the stop signal earlier. ...

TL;DR for operators Moving inactive MoE experts to SSD can make an oversized sparse model fit when the full checkpoint does not fit comfortably in DRAM. But every expert missing from fast memory must be fetched before decoding can continue. In FlashMoE’s evaluated system, an expert load takes roughly 3 milliseconds and expert loading accounts for more than 70% of decoding time. SSD therefore solves capacity only if the cache avoids enough expensive misses. ...