arXiv subjects

Keith Ross

Publications and source records attributed to Keith Ross.

26 records · Page 2Linked to original sources

arXiv2021

Randomized Ensembled Double Q-Learning: Learning Fast Without a Model

Using a high Update-To-Data (UTD) ratio, model-based methods have recently achieved much higher sample efficiency than previous model-free methods for continuous-action DRL benchmarks. In this paper, we introduce a simple model-free algorithm, Randomized Ensembled Double Q-Learning (REDQ), and show that its performance is just as good as, if not better than, a state-of-the-art model-based algorithm for the MuJoCo benchmark. Moreover, REDQ can achieve this performance using fewer parameters than the model-based method, and with less wall-clock run time. REDQ has three carefully integrated ingredients which allow it to achieve its high performance: (i) a UTD ratio >> 1; (ii) an ensemble of Q functions; (iii) in-target minimization across a random subset of Q functions from the ensemble. Through carefully designed experiments, we provide a detailed analysis of REDQ and related model-free algorithms. To our knowledge, REDQ is the first successful model-free DRL algorithm for continuous-action spaces using a UTD ratio >> 1.

cs.LG↗

arXiv2020

Striving for Simplicity and Performance in Off-Policy DRL: Output Normalization and Non-Uniform Sampling

We aim to develop off-policy DRL algorithms that not only exceed state-of-the-art performance but are also simple and minimalistic. For standard continuous control benchmarks, Soft Actor-Critic (SAC), which employs entropy maximization, currently provides state-of-the-art performance. We first demonstrate that the entropy term in SAC addresses action saturation due to the bounded nature of the action spaces, with this insight, we propose a streamlined algorithm with a simple normalization scheme or with inverted gradients. We show that both approaches can match SAC's sample efficiency performance without the need of entropy maximization, we then propose a simple non-uniform sampling method for selecting transitions from the replay buffer during training. Extensive experimental results demonstrate that our proposed sampling scheme leads to state of the art sample efficiency on challenging continuous control tasks. We combine all of our findings into one simple algorithm, which we call Streamlined Off Policy with Emphasizing Recent Experience, for which we provide robust public-domain code.

cs.LG↗

arXiv2020

BAIL: Best-Action Imitation Learning for Batch Deep Reinforcement Learning

There has recently been a surge in research in batch Deep Reinforcement Learning (DRL), which aims for learning a high-performing policy from a given dataset without additional interactions with the environment. We propose a new algorithm, Best-Action Imitation Learning (BAIL), which strives for both simplicity and performance. BAIL learns a V function, uses the V function to select actions it believes to be high-performing, and then uses those actions to train a policy network using imitation learning. For the MuJoCo benchmark, we provide a comprehensive experimental study of BAIL, comparing its performance to four other batch Q-learning and imitation-learning schemes for a large variety of batch datasets. Our experiments show that BAIL's performance is much higher than the other schemes, and is also computationally much faster than the batch Q-learning schemes.

cs.LG↗

arXiv2020

Multi-task Batch Reinforcement Learning with Metric Learning

We tackle the Multi-task Batch Reinforcement Learning problem. Given multiple datasets collected from different tasks, we train a multi-task policy to perform well in unseen tasks sampled from the same distribution. The task identities of the unseen tasks are not provided. To perform well, the policy must infer the task identity from collected transitions by modelling its dependency on states, actions and rewards. Because the different datasets may have state-action distributions with large divergence, the task inference module can learn to ignore the rewards and spuriously correlate $\textit{only}$ state-action pairs to the task identity, leading to poor test time performance. To robustify task inference, we propose a novel application of the triplet loss. To mine hard negative examples, we relabel the transitions from the training tasks by approximating their reward functions. When we allow further training on the unseen tasks, using the trained policy as an initialization leads to significantly faster convergence compared to randomly initialized policies (up to $80\%$ improvement and across 5 different Mujoco task distributions). We name our method $\textbf{MBML}$ ($\textbf{M}\text{ulti-task}$ $\textbf{B}\text{atch}$ RL with $\textbf{M}\text{etric}$ $\textbf{L}\text{earning}$).

cs.LG↗

arXiv2019

Sensing the Chinese Diaspora: How Mobile Apps Can Provide Insights into Global Migration Flows

Many countries today have "country-centric mobile apps" which are mobile apps that are primarily used by residents of a specific country. Many of these country-centric apps also include a location-based service which takes advantage of the smartphone's API access to the smartphone's current GPS location. In this paper, we investigate how such country-centric apps with location-based services can be employed to study the diaspora associated with ethnic and cultural groups. Our methodology combines GPS hacking, automated task tools for mobile phones, and OCR to generate migration statistics for diaspora. As a case study, we apply our methodology to WeChat, an enormously popular app within China and among ethnic Chinese worldwide. Using WeChat, we collect data about the Chinese diaspora in 32 cities. The combined data provides interesting insights to the modern Chinese diaspora and how it has changed in recent years.

cs.SI↗

arXiv2019

Boosting Soft Actor-Critic: Emphasizing Recent Experience without Forgetting the Past

Soft Actor-Critic (SAC) is an off-policy actor-critic deep reinforcement learning (DRL) algorithm based on maximum entropy reinforcement learning. By combining off-policy updates with an actor-critic formulation, SAC achieves state-of-the-art performance on a range of continuous-action benchmark tasks, outperforming prior on-policy and off-policy methods. The off-policy method employed by SAC samples data uniformly from past experience when performing parameter updates. We propose Emphasizing Recent Experience (ERE), a simple but powerful off-policy sampling technique, which emphasizes recently observed data while not forgetting the past. The ERE algorithm samples more aggressively from recent experience, and also orders the updates to ensure that updates from old data do not overwrite updates from new data. We compare vanilla SAC and SAC+ERE, and show that ERE is more sample efficient than vanilla SAC for continuous-action Mujoco tasks. We also consider combining SAC with Priority Experience Replay (PER), a scheme originally proposed for deep Q-learning which prioritizes the data based on temporal-difference (TD) error. We show that SAC+PER can marginally improve the sample efficiency performance of SAC, but much less so than SAC+ERE. Finally, we propose an algorithm which integrates ERE and PER and show that this hybrid algorithm can give the best results for some of the Mujoco tasks.

cs.LG↗

arXiv2017

Smoke Screener or Straight Shooter: Detecting Elite Sybil Attacks in User-Review Social Networks

Popular User-Review Social Networks (URSNs)---such as Dianping, Yelp, and Amazon---are often the targets of reputation attacks in which fake reviews are posted in order to boost or diminish the ratings of listed products and services. These attacks often emanate from a collection of accounts, called Sybils, which are collectively managed by a group of real users. A new advanced scheme, which we term elite Sybil attacks, recruits organically highly-rated accounts to generate seemingly-trustworthy and realistic-looking reviews. These elite Sybil accounts taken together form a large-scale sparsely-knit Sybil network for which existing Sybil fake-review defense systems are unlikely to succeed. In this paper, we conduct the first study to define, characterize, and detect elite Sybil attacks. We show that contemporary elite Sybil attacks have a hybrid architecture, with the first tier recruiting elite Sybil workers and distributing tasks by Sybil organizers, and with the second tier posting fake reviews for profit by elite Sybil workers. We design ElsieDet, a three-stage Sybil detection scheme, which first separates out suspicious groups of users, then identifies the campaign windows, and finally identifies elite Sybil users participating in the campaigns. We perform a large-scale empirical study on ten million reviews from Dianping, by far the most popular URSN service in China. Our results show that reviews from elite Sybil users are more spread out temporally, craft more convincing reviews, and have higher filter bypass rates. We also measure the impact of Sybil campaigns on various industries (such as cinemas, hotels, restaurants) as well as chain stores, and demonstrate that monitoring elite Sybil users over time can provide valuable early alerts against Sybil campaigns.

cs.SI↗

arXiv2008

Characterizing Video Responses in Social Networks

Video sharing sites, such as YouTube, use video responses to enhance the social interactions among their users. The video response feature allows users to interact and converse through video, by creating a video sequence that begins with an opening video and followed by video responses from other users. Our characterization is over 3.4 million videos and 400,000 video responses collected from YouTube during a 7-day period. We first analyze the characteristics of the video responses, such as popularity, duration, and geography. We then examine the social networks that emerge from the video response interactions.

cs.MM↗