Deep Neural Networks (DNNs) have shown significant advantages in a wide variety of domains. However, DNNs are becoming computationally intensive and energy hungry at an exponential pace, while at the same time, there is a vast demand for running sophisticated DNN-based services on resource constrained embedded devices. In this paper, we target energy-efficient inference on embedded DNN accelerators. To that end, we propose an automated framework to compress DNNs in a hardware-aware manner by jointly employing pruning and quantization. We explore, for the first time, per-layer fine- and coarse-grained pruning, in the same DNN architecture, in addition to low bit-width mixed-precision quantization for weights and activations. Reinforcement Learning (RL) is used to explore the associated design space and identify the pruning-quantization configuration so that the energy consumption is minimized whilst the prediction accuracy loss is retained at acceptable levels. Using our novel composite RL agent we are able to extract energy-efficient solutions without requiring retraining and/or fine tuning. Our extensive experimental evaluation over widely used DNNs and the CIFAR-10/100 and ImageNet datasets demonstrates that our framework achieves 39% average energy reduction for 1.7% average accuracy loss and outperforms significantly the state-of-the-art approaches.

The rise of Deep Neural Networks (DNNs) has resulted in complex workloads employing multiple DNNs concurrently. This trend introduces unique challenges related to workload distribution, particularly in heterogeneous embedded systems. Current run-time managers struggle to efficiently utilize all computing components on these platforms, resulting in two major problems. First, the system throughput deteriorates due to contention on the computing resources. Second, not all DNNs are affected equally, leading to inconsistent performance levels across different models. To address these challenges, we introduce FairBoost, a framework for efficient and fair multi-DNN inference on heterogeneous embedded systems. FairBoost employs Reinforcement Learning (RL) to efficiently manage multi-DNN workloads. Additionally, it incorporates a novel numerical representation of DNN layers via a Vector Quantized Variational Auto-Encoder (VQ-VAE). Finally, it enables knowledge transfer to similar heterogeneous embedded systems without retraining and/or fine-tuning. Experimental evaluation of FairBoost over 18 DNNs and various multi-DNN scenarios shows an average throughput/fairness improvement of ×3.24. Additionally, FairBoost facilitates knowledge transfer from the initial platform, Orange Pi 5, to a new system, Odroid N2+, without any retraining or fine-tuning achieving similar gains.

Pluripotent stem cells can be differentiated into all three germ-layers including ecto-, endo-, and mesoderm in vitro. However, the early identification and rapid characterization of each germ-layer in response to chemical and physical induction of differentiation is limited. This is a long-standing issue for rapid and high-throughput screening to determine lineage specification efficiency. Here, we present deep learning (DL) methodologies for predicting and classifying early mesoderm cells differentiated from embryoid bodies (EBs) based on cellular and nuclear morphologies. Using a transgenic murine embryonic stem cell (mESC) line, namely OGTR1, we validated the upregulation of mesodermal genes (Brachyury (T): DsRed) in cells derived from EBs for the deep learning model training. Cells were classified into mesodermal and non-mesodermal (representing endo- and ectoderm) classes using a convolutional neural network (CNN) model called InceptionV3 which achieved a very high classification accuracy of 97% for phase images and 90% for nuclei images. In addition, we also performed image segmentation using an Attention U-Net CNN and obtained a mean intersection over union of 61% and 69% for phase-contrast and nuclear images, respectively. This work highlights the potential of integrating cell culture, imaging technologies, and deep learning methodologies in identifying lineage specification, thus contributing to the advancements in regenerative medicine. Collectively, our trained deep learning models can predict the mesoderm cells with high accuracy based on cellular and nuclear morphologies.

We present GeoLLM-Squad, a geospatial Copilot that introduces the novel multi-agent paradigm to remote sensing (RS) workflows. Unlike existing single-agent approaches that rely on monolithic large language models (LLM), GeoLLM-Squad separates agentic orchestration from geospatial task-solving, by delegating RS tasks to specialized sub-agents. Built on the open-source AutoGen and GeoLLM-Engine frameworks, our work enables the modular integration of diverse applications, spanning urban monitoring, forestry protection, climate analysis, and agriculture studies. Our results demonstrate that while single-agent systems struggle to scale with increasing RS task complexity, GeoLLM-Squad maintains robust performance, achieving a 17% improvement in agentic correctness over state-of-the-art baselines. Our findings highlight the potential of multi-agent AI in advancing RS workflows.

Modern deep learning agents usually operate in low-dimensional environments. They process pixel input, don’t offer insights into their thought process, and require significant power and computational resources. These characteristics make them inapplicable for embedded devices. In this letter, we present Pythia, an edge-first framework that uses latent imagination to handle complex environments efficiently and envision future agent states. It utilizes a VQ-VAE to reduce the high-dimensional features into a low-dimensional space, making it ideal for modern embedded devices. Moreover, Pythia offers human interpretable feedback and scales well with respect to the design space. Pythia surpassed the other state-of-art models in prediction accuracy on both intrinsic and extrinsic metrics.