We introduce SuperCLUE-Math6(SC-Math6), a new benchmark dataset to evaluate the mathematical reasoning abilities of Chinese language models. SC-Math6 is designed as an upgraded Chinese version of the GSM8K dataset with enhanced difficulty, diversity, and application scope. It consists of over 2000 mathematical word problems requiring multi-step reasoning and providing natural language solutions. We propose an innovative scheme to quantify the reasoning capability of large models based on performance over problems with different reasoning steps. Experiments on 13 representative Chinese models demonstrate a clear stratification of reasoning levels, with top models like GPT-4 showing superior performance. SC-Math6 fills the gap in Chinese mathematical reasoning benchmarks and provides a comprehensive testbed to advance the intelligence of Chinese language models.

Sequential incentive marketing is an important approach for online businesses to acquire customers, increase loyalty and boost sales. How to effectively allocate the incentives so as to maximize the return (e.g., business objectives) under the budget constraint, however, is less studied in the literature. This problem is technically challenging due to the facts that 1) the allocation strategy has to be learned using historically logged data, which is counterfactual in nature, and 2) both the optimality and feasibility (i.e., that cost cannot exceed budget) needs to be assessed before being deployed to online systems. In this paper, we formulate the problem as a constrained Markov decision process (CMDP). To solve the CMDP problem with logged counterfactual data, we propose an efficient learning algorithm which combines bisection search and model-based planning. First, the CMDP is converted into its dual using Lagrangian relaxation, which is proved to be monotonic with respect to the dual variable. Furthermore, we show that the dual problem can be solved by policy learning, with the optimal dual variable being found efficiently via bisection search (i.e., by taking advantage of the monotonicity). Lastly, we show that model-based planing can be used to effectively accelerate the joint optimization process without retraining the policy for every dual variable. Empirical results on synthetic and real marketing datasets confirm the effectiveness of our methods.

As he entered Hong Kong Supermarket, Sam Lin scanned text messages from his wife instructing him how many red envelopes to buy. Three dozen, she wrote -- and make them large, to fit checks rather than folded wads of cash. Lin's nephews, nieces and in-laws will not have the thrill of pulling crisp bills out of their red Lunar New Year good luck envelopes when the Year of the Ox begins Friday. Normally, Lin goes to his credit union weeks ahead of the holiday to pre-order new bills -- a total of $900 to $1,000 for the kids and elders in his extended family. But with the possibility that the coronavirus could be lurking on $20 or $100 bills, Lin is one of many Asian Americans forgoing traditional cash to ring in the festivities.

Prediction 1: Superhero CMOs Emerge - By 2020, the first superhero CMOs will emerge because they received C-Level permission to disrupt traditional go-to-market operations. IDC predicts that we will see pockets of break-through in CMO leadership. This will be demonstrated by individuals who have exceptional leadership skills and in the face of long-odds, have brought meaningful change to their marketing organizations. The "Superhero CMO" will be one who has executed real change -- not just the aspirational change that is depicted on a PowerPoint slide. Prediction 2: Boardroom Battle for the Customer - By 2020, 25% of CEO's will appoint a Chief Customer Officer (CCO) in an attempt to unify the imperative of customer-centricity.

The media has declared chatbots the digital version of the little black dress: a technology staple that every brand must now have and every payment type must now commerce-enable. Technology enthusiasts draw pictures on whiteboards showing bots as the new king of the stack, out-stacking the last great stack -- apps, which sit on top of the next-to-last great stack -- mobile operating systems, which sit on top of the next-to-next-to-last great stack -- the web. VCs tout bots as commerce's next big frontier, one that can make the notion of contextual, conversational commerce real inside of the app where consumers now spend 75 percent of their time: messaging. The big bet is that bots will make it possible for brands to indulge the every digital whim of this group of consumers without them ever having to leave their digital home away from home. Facebook Messenger's announcement last week that 30,000 bots would soon be payment-enabled for the 900 million people who hang inside of its messaging ecosystem only stoked that fire.