The Chinese University of Hong Kong

About: The Chinese University of Hong Kong is a education organization based out in Hong Kong, China. It is known for research contribution in the topics: Population & Computer science. The organization has 43411 authors who have published 93672 publications receiving 3066651 citations.

Towards Diverse and Natural Image Descriptions via a Conditional GAN

Abstract: Despite the substantial progress in recent years, the image captioning techniques are still far from being perfect. Sentences produced by existing methods, e.g. those based on RNNs, are often overly rigid and lacking in variability. This issue is related to a learning principle widely used in practice, that is, to maximize the likelihood of training samples. This principle encourages high resemblance to the “ground-truth” captions, while suppressing other reasonable descriptions. Conventional evaluation metrics, e.g. BLEU and METEOR, also favor such restrictive methods. In this paper, we explore an alternative approach, with the aim to improve the naturalness and diversity – two essential properties of human expression. Specifically, we propose a new framework based on Conditional Generative Adversarial Networks (CGAN), which jointly learns a generator to produce descriptions conditioned on images and an evaluator to assess how well a description fits the visual content. It is noteworthy that training a sequence generator is nontrivial. We overcome the difficulty by Policy Gradient, a strategy stemming from Reinforcement Learning, which allows the generator to receive early feedback along the way. We tested our method on two large datasets, where it performed competitively against real people in our user study and outperformed other methods on various tasks.

Innovation and Growth: How Business Contributes to Society

Abstract: Executive Overview Milton Friedman once argued that profits are the chief purpose of business. Profits do matter, but today we know more about how business contributes to society. Good firms bring innovation to the marketplace, which facilitates their growth. Innovative, growing firms generate economic growth and employment, which, in turn, greatly improves people's lives. In this paper I argue that the main goal of business is to develop new and innovative goods and services that generate economic growth while delivering important benefits to society. Steady economic growth generated through innovation plays a major role in producing increases in per capita income. Small changes in economic growth can yield very large differences in income over time, making firm growth particularly salient to societies. In addition to providing growth, innovative firms can supply important goods and services to consumers, particularly those at the base of the pyramid. Through innovation and growth firms can do untold goo...

Compression Artifacts Reduction by a Deep Convolutional Network

Abstract: Lossy compression introduces complex compression artifacts, particularly the blocking artifacts, ringing effects and blurring. Existing algorithms either focus on removing blocking artifacts and produce blurred output, or restores sharpened images that are accompanied with ringing effects. Inspired by the deep convolutional networks (DCN) on super-resolution, we formulate a compact and efficient network for seamless attenuation of different compression artifacts. We also demonstrate that a deeper model can be effectively trained with the features learned in a shallow network. Following a similar "easy to hard" idea, we systematically investigate several practical transfer settings and show the effectiveness of transfer learning in low-level vision problems. Our method shows superior performance than the state-of-the-arts both on the benchmark datasets and the real-world use case (i.e. Twitter). In addition, we show that our method can be applied as pre-processing to facilitate other low-level vision routines when they take compressed images as input.

Detection and Dynamic Changes of EGFR Mutations from Circulating Tumor DNA as a Predictor of Survival Outcomes in NSCLC Patients Treated with First-line Intercalated Erlotinib and Chemotherapy

Abstract: Purpose: Blood-based circulating-free (cf) tumor DNA may be an alternative to tissue-based EGFR mutation testing in NSCLC. This exploratory analysis compares matched tumor and blood samples from the FASTACT-2 study. Experimental Design: Patients were randomized to receive six cycles of gemcitabine/platinum plus sequential erlotinib or placebo. EGFR mutation testing was performed using the cobas tissue test and the cobas blood test (in development). Blood samples at baseline, cycle 3, and progression were assessed for blood test detection rate, sensitivity, and specificity; concordance with matched tumor analysis ( n = 238), and correlation with progression-free survival (PFS) and overall survival (OS). Results: Concordance between tissue and blood tests was 88%, with blood test sensitivity of 75% and a specificity of 96%. Median PFS was 13.1 versus 6.0 months for erlotinib and placebo, respectively, for those with baseline EGFR mut + cfDNA [HR, 0.22; 95% confidence intervals (CI), 0.14–0.33, P EGFR mut − cfDNA subgroup (HR, 0.83; 95% CI, 0.65–1.04, P = 0.1076). For patients with EGFR mut + cfDNA at baseline, median PFS was 7.2 versus 12.0 months for cycle 3 EGFR mut + cfDNA versus cycle 3 EGFR mut − patients, respectively (HR, 0.32; 95% CI, 0.21–0.48, P P = 0.0066). Conclusions: Blood-based EGFR mutation analysis is relatively sensitive and highly specific. Dynamic changes in cfDNA EGFR mutation status relative to baseline may predict clinical outcomes. Clin Cancer Res; 21(14); 3196–203. ©2015 AACR .