Internet has witnessed its paramount function transition from hostto-host communication to content dissemination. Named Data Networking (NDN) and Content-Centric Networking (CCN) emerge as a clean slate network architecture to embrace this shift. Pending Interest Table (PIT) in NDN/CCN keeps track of the Interest packets that are received but yet un-responded, which brings NDN/CCN significant features, such as communicating without the knowledge of source or destination, loop and packet loss detection, multipath routing, better security, etc.This paper presents a thorough study of PIT for the first time. Using an approximate, application-driven translation of current IPgenerated trace to NDN trace, we firstly quantify the size and access frequencies of PIT. Evaluation results on a 20 Gbps gateway trace show that the corresponding PIT contains 1.5 M entries, and the lookup, insert and delete frequencies are 1.4 M/s, 0.9 M/s and 0.9 M/s, respectively. Faced with this challenging issue and to make PIT more scalable, we further propose a Name Component Encoding (NCE) solution to shrink PIT size and accelerate PIT access operations. By NCE, the memory consumption can be reduced by up to 87.44%, and the access performance significantly advanced, satisfying the access speed required by PIT. Moreover, PIT exhibits good scalability with NCE. At last, we propose to place PIT on (egress channel of) the outgoing line-cards of routers, which meets the NDN design and eliminates the cumbersome synchronization problem among multiple PITs on the line-cards.

No abstract

No abstract

A common vision from science fiction is that robots will one day inhabit our physical spaces, sense the world as we do, assist our physical labours, and communicate with us through natural language. Here we study how to design artificial agents that can interact naturally with humans using the simplification of a virtual environment. This setting nevertheless integrates a number of the central challenges of artificial intelligence (AI) research: complex visual perception and goal-directed physical control, grounded language comprehension and production, and multi-agent social interaction. To build agents that can robustly interact with humans, we would ideally train them while they interact with humans. However, this is presently impractical. Therefore, we approximate the role of the human with another learned agent, and use ideas from inverse reinforcement learning to reduce the disparities between human-human and agent-agent interactive behaviour. Rigorously evaluating our agents poses a great challenge, so we develop a variety of behavioural tests, including evaluation by humans who watch videos of agents or interact directly with them. These evaluations convincingly demonstrate that interactive training and auxiliary losses improve agent behaviour beyond what is achieved by supervised learning of actions alone. Further, we demonstrate that agent capabilities generalise beyond literal experiences in the dataset. Finally, we train evaluation models whose ratings of agents agree well with human judgement, thus permitting the evaluation of new agent models without additional effort. Taken together, our results in this virtual environment provide evidence that large-scale human behavioural imitation is a promising tool to create intelligent, interactive agents, and the challenge of reliably evaluating such agents is possible to surmount. See videos for an overview of the manuscript, training time-lapse, and human-agent interactions.

No abstract