Cold Start

Polymer electrolyte membrane fuel cell (PEMFC) has been recognized as a promising zero-emission power source for portable, mobile and stationary applications. To simultaneously ensure high membrane proton conductivity and sufficient reactant delivery to reaction sites, water management has become one of the most important issues for PEMFC commercialization, and proper water management requires good understanding of water transport in different components of PEMFC. In this paper, previous researches related to water transport in PEMFC are comprehensively reviewed. The state and transport mechanism of water in different components are elaborated in detail. Based on the literature review, it is found that experimental techniques have been developed to predict distributions of water, gas species, temperature and other parameters in PEMFC. However, difficulties still remain for simultaneous measurements of multiple parameters, and the cell and system design modifications required by measurements need to be minimized. Previous modeling work on water transport in PEMFC involves developing rule-based and first-principle-based models, and first-principle-based models involve multi-scale methods from atomistic to full cell levels. Different models have been adopted for different purposes and they all together can provide a comprehensive view of water transport in PEMFC. With the development of computational power, application of lower length scale methods to higher length scales for more accurate and comprehensive results is feasible in the future. Researches related to cold start (startup from subzero temperatures) and high temperature PEMFC (HT-PEMFC) (operating at the temperatures higher than 100 C) are also reviewed. Ice formation that hinders reactant delivery and damages cell materials is the major issue for PEMFC cold start, and enhancing water absorption by membrane electrolyte and external heating have been identified as the most effective ways to reduce ice formation and accelerate temperature increment. HT-PEMFC that can operate without liquid water formation and membrane hydration greatly simplifies water management strategy, and promising performance of HT-PEMFC has been demonstrated.

Recommendation systems are essential tools to overcome the choice overload problem by suggesting items of interest to users. However, they suffer from a major challenge which is the so-called cold-start problem. The cold-start problem typically happens when the system does not have any form of data on new users and on new items. In this chapter, we describe the cold start problem in recommendation systems. We mainly focus on Collaborative Filtering (CF) systems which are the most popular approaches to build recommender systems and have been successfully employed in many real-world applications. Moreover, we discuss multiple scenarios that cold-start may happen in these systems and explain different solutions for them.

In the ecommerce services, there is a very important tool that will determine the success to increase number of buying and selling in the marketing target, that is how the user in finding products that are suitable to be purchased. This tool is called recommender system. Recommender system is important tool for establishing an effective communication between users and retailers in ecommerce business. Effective and enjoyable communication to find the product is considered to have a significant impact that increase of marketing achievement. Recommender system established in the mid-90s. Based on technical approach, there are four types of recommender system namely Collaborative, Contents Based, Knowledge Based and Demographic filtering. Collaborative filtering is considered to be more superior than another two methods. It offers obviously advantages in terms of serendipity, novelty and accuracy. Although it has some benefit. However, there is a critical problem in collaborative filtering that called cold start. It is a major problem to which many researchers have paid much more attention to this particular research interest. In response to this particular problem the critical review and analysis on state of the art of the current technology, some possible solutions including approach, method and techniques used have been identified but they need further validation.

The increasing growth of the World Wide Web especially in a social network with the multiplicity of items offered (such as products or web pages), it is really difficult for a user to pick up relevant items who is searching for it. On the other hand, the different tastes and behaviors of users is making the probability for finding a neighbor user hard to get, therefore, difficult for automated software systems to discover what is interesting to the user. We have proposed a new approach to adapt to this widespread in e-commerce nowadays and reduce the impact of the multiplicity of items and different views of the users that can quickly produce the recommendations through, exploit the domain knowledge of training data set to create testing data set depending on attributes of one feature that represents the distinctive genres of item as the inputs to a hybrid recommender systems which is aspired to achieve best recommendations by implementing meta-level hybridization techniques that combine of collaborative recommender systems and content-based recommender systems, these operations will reduce from the effects of sparsity, cold start and scalability very common problems with the collaborative recommender systems additional to improve the accuracy of recommendation comparing with the pure collaborative filtering Pearson Correlation approach.

A prerequisite for implementing collaborative filtering recommender systems is the availability of users' preferences data. This data, typically in the form of ratings, is exploited to learn the tastes of the users and to serve them with personalized recommendations. However, there may be a lack of preference data, especially at the initial stage of the operations of a recommender system, i.e., in the Cold Start phase. In particular, when a new user has not yet rated any item, the system would be incapable of generating relevant recommendations for this user. Or, when a new item is added to the system catalogue and no user has rated it, the system cannot recommend this item to any user. This chapter discusses the cold start problem and provides a comprehensive description of techniques that have been proposed to address this problem. It surveys algorithmic solutions and provides a summary of their performance comparison. Moreover, it lists publicly available resources (e.g., libraries and datasets) and offers a set of practical guidelines that can be adopted by researchers and practitioners.

Item features play an important role in movie recommender systems, where recommendations can be generated by using explicit or implicit preferences of users on traditional features (attributes) such as tag, genre, and cast. Typically, movie features are human-generated, either editorially (e.g., genre and cast) or by leveraging the wisdom of the crowd (e.g., tag), and as such, they are prone to noise and are expensive to collect. Moreover, these features are often rare or absent for new items, making it difficult or even impossible to provide good quality recommendations. In this paper, we show that users’ preferences on movies can be well or even better described in terms of the mise-en-scène features, i.e., the visual aspects of a movie that characterize design, aesthetics and style (e.g., colors, textures). We use both MPEG-7 visual descriptors and Deep Learning hidden layers as examples of mise-en-scène features that can visually describe movies. These features can be computed automatically from any video file, offering the flexibility in handling new items, avoiding the need for costly and error-prone human-based tagging, and providing good scalability. We have conducted a set of experiments on a large catalog of 4K movies. Results show that recommendations based on mise-en-scène features consistently outperform traditional metadata attributes (e.g., genre and tag).

Music recommender systems (MRSs) have experienced a boom in recent years, thanks to the emergence and success of online streaming services, which nowadays make available almost all music in the world at the user's fingertip. While today's MRSs considerably help users to find interesting music in these huge catalogs, MRS research is still facing substantial challenges. In particular when it comes to build, incorporate, and evaluate recommendation strategies that integrate information beyond simple user-item interactions or content-based descriptors, but dig deep into the very essence of listener needs, preferences, and intentions, MRS research becomes a big endeavor and related publications quite sparse. The purpose of this trends and survey article is twofold. We first identify and shed light on what we believe are the most pressing challenges MRS research is facing, from both academic and industry perspectives. We review the state of the art toward solving these challenges and discuss its limitations. Second, we detail possible future directions and visions we contemplate for the further evolution of the field. The article should therefore serve two purposes: giving the interested reader an overview of current challenges in MRS research and providing guidance for young researchers by identifying interesting, yet under-researched, directions in the field.

Recent studies relevant to the future use of gold catalysts in the automotive industry are summarized. Gold catalysts have been examined for their potential in low temperature activity to combat cold-start emission problems and removal of NOx from leanburn gasoline and diesel engines. The justification for developing gold catalyst technologies is based both on their promising technical performance and the relatively stable price and greater availability of gold compared with the platinum group metals (PGMs). Use of gold catalysts under mild conditions could also produce lower proportions of CO in the hydrogen streams used for automotive fuel cells and they could be used for the selective oxidation of carbon monoxide in these hydrogen streams in order to increase the efficiency of these fuel cells. Technical barriers are indicated which still remain to be overcome before application of gold catalysts is successful in the automotive sphere.

Recommender Systems (RS) are software tools which can be used in making useful predictions of items to users. RS has been an important research area since the mid-1990s, and there are a lot of RS tools built since then to improve user satisfaction. While building the RS tools, researchers face a lot of problems in the form of data sparsity, information overload, cold-start, scalability, lack of resources and time. These factors may reduce the accuracy of predictions. To overcome these problems, researchers model the rating data as graphs. Through graphs, we can explore the transitive associations and hidden information in our dataset. Particularly, this work focuses only on the graph-based collaborative filtering (GBCF) techniques introduced in the recommendation systems. We have studied and analyzed various GBCF articles published in the most popular online digital libraries during the last two decades. These approaches have been categorized into three broader categories: user-item, user-user, and item-item based on the model of the graph, which is used for the recommendation purpose. This survey provides an understanding of how graph-based CF has helped researchers and developers built a more efficient RS model in terms of different RS goals, such as accuracy.

This study investigated the phenomenon of water freezing below freezing point in polymer electrolyte fuel cells (PEFCs). To understand the details of water freezing phenomena inside a PEFC, a system capable of cross-sectional imaging inside the fuel cell with visible and infrared images was developed. Super-cooled water freezing phenomena were observed under different gas purge conditions. The present test confirmed that super-cooled water was generated on the gas diffusion layer (GDL) surface and that water freezing occurs at the interface between the GDL and MEA (membrane electrode assembly) at the moment cell performance deteriorates under conditions when remaining water was dry enough inside the fuel cell before cold starting. Moreover, using infrared radiation imaging, it was clarified that heat of solidification spreads at the GDL/MEA interface at the moment cell performance drops. Compared with a no-initial purge condition, liquid water generation was not confirmed to cause ice growth at the GDL/MEA interface after cell performance deterioration. Each condition indicated that ice formation at the GDL/MEA interface causes cell performance deterioration. Therefore, it is believed that ice formation between the GDL/MEA interface causes air gas stoppage and that this blockage leads to a drop in cell performance.

Collaborative filtering is one of the most common approaches in many current recommender systems. However, historical data and customer profileles, necessary for this approach, are not always available. Similarly, new products are constantly launched to the market lacking historical information. We propose a new method to deal with these "cold start" scenarios, designing price-estimation functions used for making recommendations based on cost-benefit analysis. Experimental results, using a data set of 836 laptop descriptions, showed that such price-estimation functions can be learned from data. Besides, they can also be used to formulate interpretable recommendations that explain to users how product features determine its price. Finally a 2D visualization of the proposed recommender system was provided.

In this paper we evaluate the performance of different collaborative algorithms in cold-start situations, where the initial lack of ratings may affect the quality of the algorithms. Instead of trying to design an algorithm that provides, on average, a reasonable good accuracy in both the early and steady stages of a recommender system, in this paper we suggest adopting different algorithms in the different stages of the recommender system life-cycle.

Product recommendations are generally understood as data-drivenhowever, we argue that knowledge-driven management decisions may also play a role, especially in the cold start problem, which has been tackled with various degrees of success by different techniques. We, therefore, advocate an approach based on capturing managerial priorities in the act of recommendation building-our proposal is to complement the customer-centric view (affected by uncertainty) with a machine-readable business-centric view. For this purpose, the paper presents a computational pipeline for the "digital nudging" of recommendations-it starts by capturing a manager's priorities with diagrammatic means, which are further exposed as a Knowledge Graph to a recommender built on a modified version of the Onicescu method, which takes into consideration a business utility concept to inform decision-making. The research follows the Design Science methodology, resulting in an artifact that tackles the cold-start through a bydesign "digital nudging" mechanism embedded in the recommender system through a semantic pipeline whose ingredients are (a) Agile Modeling Method Engineering (to setup up the diagrammatic tool for prioritization rules), (b) the RDF framework (to capture the diagrammatic rules in a knowledge graph), and (c) the Onicescu multi-criteria decision method modified according to Zipf's Law, made popular by Internet marketing. The evaluation was based on surveys with potential stakeholders. The implications are that the notion of "digital nudging" can take a knowledge-driven form, engineered as an artifact that bridges the decision-makers' perspective and priorities (captured by diagrammatic means) with the customerfacing output (recommendations), instead of relying solely on the accumulated history of transactional data. This interpretation of digital nudging may be extended towards other "digital choice environments" where contextual decisions are called to influence computational output. Limitations are related to the experimental nature of the implementation tools (proof-of-concept), where some manual graph transformations are still needed to pass the information through the pipeline.

The lack of treatment capacity of hydrocarbons by three-way catalysts during the “cold start” period creates an important environmental problem. During this period, the temperature of the three-way catalyst is too low for effective operation and cannot convert the hydrocarbons in the exhaust. 50–80% of the total hydrocarbon emissions are produced in this phase that accomplishes the first 60–120 s of the engine operation. In this study, the technology chosen to treat these emissions is the use of HC-traps, and molecular simulations are tested as a tool to reproduce the experimental adsorption behaviour of porous solids. Therefore, experimental and simulated adsorption isotherms of propene (model hydrocarbon) have been obtained for four different crystalline materials with distinctive framework structures (3D and 1D) and a variety of Si/Al ratios and cations (three zeolites: ZSM-5, BETA and Mordenite; and a silicoaluminophosphate molecular sieve: SAPO-5).

The exploitation of an SOFC-system model to define and test control and energy management strategies is presented. Such a work is motivated by the increasing interest paid to SOFC technology by industries and governments due to its highly appealing potentialities in terms of energy savings, fuel flexibility, cogeneration, low-pollution and low-noise operation.

In a two-level computer storage hierarchy, miss ratio measurements are often made from a "cold start", that is, made with the first-level store initially empty. For large capacities the effect on the measured miss ratio of the misses incurred while filling the first-level store can be significant, even for long reference strings. Use of "warm-start" rather than "cold-start" miss ratios cast doubt on the widespread belief that the observed "S-shape" of lifetime (reciprocal of miss ratio) versus capacity curve indicates a property of behavior of programs that maintain a constant number of pages in main storage. On the other hand, if cold-start miss ratios are measured as a function of capacity and measurement length, then they are useful in studying systems in which operation of a program is periodically interrupted by task switches. It is shown how to obtain, under simple assumptions, the cache miss ratio for multiprogramming from cold-start miss ratio values and how to obtain approximate cold-start miss ratios from warm-start miss ratios.

In this paper we present a strategy for speeding up the estimation of expected maximum flows through reliable networks. Our strategy tries to minimize the repetition of computational effort while evaluating network states sampled using the crude Monte Carlo method. Computational experiments with this strategy on three types of randomly generated networks show that it reduces the number of flow augmentations required for evaluating the states in the sample by as much as 52% on average with a standard deviation of 7% compared to the conventional strategy. This leads to an average time saving of about 71% with a standard deviation of about 8%.

In this paper we present a strategy for speeding up the estimation of expected maximum flows through reliable networks. Our strategy tries to minimize the repetition of computational effort while evaluating network states sampled using the crude Monte Carlo method. Computational experiments with this strategy on three types of randomly generated networks show that it reduces the number of flow augmentations required for evaluating the states in the sample by as much as 52% on average with a standard deviation of 7% compared to the conventional strategy. This leads to an average time saving of about 71% with a standard deviation of about 8%.

Music recommender systems (MRSs) have experienced a boom in recent years, thanks to the emergence and success of online streaming services, which nowadays make available almost all music in the world at the user's fingertip. While today's MRSs considerably help users to find interesting music in these huge catalogs, MRS research is still facing substantial challenges. In particular when it comes to build, incorporate, and evaluate recommendation strategies that integrate information beyond simple user-item interactions or content-based descriptors, but dig deep into the very essence of listener needs, preferences, and intentions, MRS research becomes a big endeavor and related publications quite sparse. The purpose of this trends and survey article is twofold. We first identify and shed light on what we believe are the most pressing challenges MRS research is facing, from both academic and industry perspectives. We review the state of the art toward solving these challenges and discuss its limitations. Second, we detail possible future directions and visions we contemplate for the further evolution of the field. The article should therefore serve two purposes: giving the interested reader an overview of current challenges in MRS research and providing guidance for young researchers by identifying interesting, yet under-researched, directions in the field. Keywords Music recommender systems • Challenges • Automatic playlist continuation • User-centric computing This research was supported in part by the Center for Intelligent Information Retrieval. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect those of the sponsors.

In the past few years, with the proliferation of mobile devices people are experiencing frequent communication and information exchange. For instance, in the context of people’s visits, it is often the case that each person carries out a smart phone, to get information about nearby places. When one visits some location, an application will recommend useful information according to its current location, preferences and past visits. This paper introduces the concept of smart routing as a recommender system for tourists that takes into account the dynamics of their personal user profiles. The concept relies on the hybrid recommender system for map routing which can be used as the alternative to the current systems. This system implements hybridization of content and collaborative-filtering. The purpose of this paper is to build a system for web and mobile based applications with the idea of recommending parameters like food, restaurants, hospitals, ATMs etc. according to users’ current location, preferences and past visits. This approach has its roots in information filtering and retrieval research. The proposed system has also taken care of the cold start problem for new users.

The ignition-extinction processes for CO oxidation over a Pt/Al 2 O 3 monolith catalyst have been studied by flow-reactor experiments and simulations. The study was performed by stepwise changes of the inlet O 2 concentration ranging 0-20 vol% while the CO concentration and the inlet gas temperature were kept constant at 1.0 vol% and 423 K, respectively. Several features observed experimentally are qualitatively simulated with our model: (i) the ignition of the CO oxidation demands 8.0 vol% O 2 (ii) corresponding to a catalyst ignition temperature of 433 K (due to the exothermicity of the reaction) and (iii) occurs in the rear part of the monolith where (iv) a local reaction zone is formed which (v) moves towards the reactor inlet as a function of time on stream. Additionally, the simulations show first order kinetic phase transitions, i.e. rapid adsorbate concentration changes, where the catalyst surface is predominantly CO covered in the low reactive state and almost completely oxygen covered in the high reactive state. For the ignition process the kinetic phase transition occurs after the actual catalytic ignition. However, the extinction process is more difficult to simulate dynamically without changing the model parameters for O 2 adsorption in the low and high reactive state, respectively. The influence of diffusion limitations and the role of formation of a less reactive Pt state under oxidising conditions is discussed.

In this paper we present a strategy for speeding up the estimation of expected maximum flows through reliable networks. Our strategy tries to minimize the repetition of computational effort while evaluating network states sampled using the crude Monte Carlo method. Computational experiments with this strategy on three types of randomly generated networks show that it reduces the number of flow augmentations required for evaluating the states in the sample by as much as 52% on average with a standard deviation of 7% compared to the conventional strategy. This leads to an average time saving of about 71% with a standard deviation of about 8%.

In this paper we present a strategy for speeding up the estimation of expected maximum flows through reliable networks. Our strategy tries to minimize the repetition of computational effort while evaluating network states sampled using the crude Monte Carlo method. Computational experiments with this strategy on three types of randomly generated networks show that it reduces the number of flow augmentations required for evaluating the states in the sample by as much as 52% on average with a standard deviation of 7% compared to the conventional strategy. This leads to an average time saving of about 71% with a standard deviation of about 8%.

In this paper we present a strategy for speeding up the estimation of expected maximum flows through reliable networks. Our strategy tries to minimize the repetition of computational effort while evaluating network states sampled using the crude Monte Carlo method. Computational experiments with this strategy on three types of randomly generated networks show that it reduces the number of flow augmentations required for evaluating the states in the sample by as much as 52% on average with a standard deviation of 7% compared to the conventional strategy. This leads to an average time saving of about 71% with a standard deviation of about 8%.

The goal of a recommender system is to suggest items of interest to a user based on historical behavior of a community of users. Given detailed enough history, item-based collaborative filtering (CF) often performs as well or better than almost any other recommendation method. However, in cold-start situations-where a user, an item, or the entire system is new-simple non-personalized recommendations often fare better. We improve the scalability and performance of a previous approach to handling cold-start situations that uses filterbots, or surrogate users that rate items based only on user or item attributes. We show that introducing a very small number of simple filterbots helps make CF algorithms more robust. In particular, adding just seven global filterbots improves both user-based and item-based CF in cold-start user, cold-start item, and cold-start system settings. Performance is better when data is scarce, performance is no worse when data is plentiful, and algorithm efficiency is negligibly affected. We systematically compare a non-personalized baseline, user-based CF, item-based CF, and our bot-augmented user-and item-based CF algorithms using three data sets (Yahoo! Movies, MovieLens, and Each-Movie) with the normalized MAE metric in three types of cold-start situations. The advantage of our "naïve filterbot" approach is most pronounced for the Yahoo! data, the sparsest of the three data sets.

In this paper we present a strategy for speeding up the estimation of expected maximum flows through reliable networks. Our strategy tries to minimize the repetition of computational effort while evaluating network states sampled using the crude Monte Carlo method. Computational experiments with this strategy on three types of randomly generated networks show that it reduces the number of flow augmentations required for evaluating the states in the sample by as much as 52% on average with a standard deviation of 7% compared to the conventional strategy. This leads to an average time saving of about 71% with a standard deviation of about 8%.

A 3-cylinder port fuel injection engine was adopted to study engine power, torque, fuel economy, emissions including regulated and non-regulated pollutants and cold start performance with the fuel of low fraction methanol in gasoline. Without any retrofit of the engine, experiments show that the engine power and torque will decrease with the increase fraction of methanol in the fuel blends under wide open throttle (WOT) conditions. However, if spark ignition timing is advanced, the engine power and torque can be improved under WOT operating conditions. Engine thermal efficiency is thus improved in almost all operating conditions. Engine combustion analyses show that the fast burning phase becomes shorter, however, the flame development phase is a little delay. When methanol/gasoline fuel blends being used, the engine emissions of carbon monoxide (CO) and hydrocarbon (HC) decrease, nitrogen oxides (NO x) changes little prior to three-way catalytic converter (TWC). After TWC, the conversion efficiencies of HC, CO and NO x are better. The non-regulated emissions, unburned methanol and formaldehyde, increase with the fraction of methanol, engine speed and load, and generally the maximum concentrations are less than 200 ppm. Experimental tests further prove that methanol and formaldehyde can be oxidized effectively by TWC. During the cold start and warming-up process at 5°C, with methanol addition into gasoline, HC and CO emissions decrease obviously. HC emission reduces more than 50% in the first few seconds (cold start period) and nearly 30% in the following warming-up period, CO reduces nearly 25% when the engine is fueled with M30. Meanwhile, the temperature of exhaust increases, which is good to activate TWC.

The cold start behaviour of the gas processing unit is one crucial issue for the use of gasoline based fuel reformers for mobile fuel cell systems. In this contribution different cold start strategies for a mobile fuel reformer based on gasoline are presented and discussed. The simulation studies are based on 1-d, dynamic multiphase models for both an autothermal gasoline reformer (ATR) and a thermally integrated reforming unit consisting of an ATR, a heat exchanger and a high-temperature-shift-reactor (HTS). Setup and geometric parameters for both models correspond to pilot stage systems considered by DaimlerChrysler. Results on the reactive heat-up of the ATR by partial and total oxidation of gasoline show the impact of the air/fuel-ratio and the thermal load on the cold start duration. The use of the reformat during the rapid start-up of the ATR is mainly limited by the availability of steam for autothermal operation. Due to the high thermal capacities of the system, the whole reforming unit requires much longer time for the cold start. Especially the slow convective heat-up of the HTS restricts the conversion of CO and the subsequent use of the reformat in the fuel cell. Several options for the acceleration of the cold start were investigated. Both a simple λ-control strategy and the reactive heat-up of the HTS by (partial) oxidation of the reformat with injected air reduce the cold start time significantly. With these measures a hydrogen-rich reformat with acceptable CO-concentration is available within two minutes. Moreover, the cold start time can be further reduced, if the HTS is heated up electrically to their ignition temperature at the beginning of the cold start. Thereby the CO-conversion in the HTS already starts in the first minute and, depending on the availability of steam for the feed stream, a cold start of the reforming unit below one minute seems to be possible.