Self-supervised learning solves pretext prediction tasks that do not require annotations to learn feature representations. For vision tasks, pretext tasks such as predicting rotation, solving jigsaw are solely created from the input data. Yet, predicting this known information helps in learning representations useful for downstream tasks. However, recent works have shown that wider and deeper models benefit more from self-supervised learning than smaller models. To address the issue of self-supervised pre-training of smaller models, we propose Distill-on-the-Go (DoGo), a self-supervised learning paradigm using single-stage online knowledge distillation to improve the representation quality of the smaller models. We employ deep mutual learning strategy in which two models collaboratively learn from each other to improve one another. Specifically, each model is trained using self-supervised learning along with distillation that aligns each model's softmax probabilities of similarity scores with that of the peer model. We conduct extensive experiments on multiple benchmark datasets, learning objectives, and architectures to demonstrate the potential of our proposed method. Our results show significant performance gain in the presence of noisy and limited labels and generalization to out-of-distribution data.

Dogo goes up stairs, rolls out in pairs, and fits on a squadmate's back. Dogo is a new remotely operated robot made by Israel's General Robotics, and inside its tiny frame it carries another, more familiar weapon. And then they reveal their deadly secret. Nestled sideways inside the Dogo is a 9mm Glock 26 compact pistol. It's like the kind of thing Nerf would make, only it's real and instead of foam darts this shoots actual bullets to kill actual people. Like the Nerf toy, Dogo is remotely controlled from a special handset, which displays live video and lets the human remotely fire the weapon.

The Dogo from General Robotics is a portable, tactical combat robot. The big difference is that the Dogo is armed with a 9mm Glock pistol. Dogo weighs 26 pounds and can to climb stairs and other obstacles. It trundles along at 2.5 miles an hour for at least two hours on one battery charge. The operator gets a 360-degree view of the surroundings via six video cameras, while another two cameras are sighted along the pistol barrel.

"Criminal was killed by robot" might soon become a standard police report. Just as an anonymous "police sniper," this robot gives the operator a chance to remain anonymous while pressing the red button to relay a kill command. The DOGO crawling traction tactical combat robot unveiled by Israeli General Robotics Ltd. was developed with counterterrorism in mind. The machine, operated remotely from a Panasonic touchpad, boasts 360-degree vision and audio intercom for conducting negotiations. It can be armed with non-lethal weapons, such as pepper spray and flashlight blinders, but it can also serve as an unemotional and lethal terminator.