Why it Matters for Data Scientists Supervised machine learning algorithms function under the assumption that the training data belongs in the same feature space and has the same distribution than the data used for inference. In fact, if a trained model fails to generalize, the very first thing a data scientist would do is check if both statistical distributions match. Wouldn’t it be amazing though if instead, there was a way to leverage data from an entirely different domain where data is abundant and cheap to label, and use it as training data for a problem that we have difficulties to solve in particular because of data sparsity? Well, this is exactly what transfer learning sets itself to achieve.As we have seen, transfer learning helps address the potential challenges related to the collection and labeling of large amounts of data that would typically be necessary for supervised learning. Beyond the promise of applying pre-learned knowledge to entirely new tasks, there is an implicit expectation from data scientists that by reusing this precedent knowledge, the model performance should improve, and the training time should dramatically decrease compared to “traditional” supervised learning scenarios.In fact, in successful cases of transfer learning, data scientists can expect the several practical benefits such as improving the performance of learning while avoiding extensive data collection and labeling efforts: August 16, 2018 at 6:26 am They hope that the final performance of the converged skill of the trained model is better than it would be without transfer learning. To answer those questions, it is worthwhile to dig deeper into the theory of transfer learning to understand the different transfer learning approaches that exist.As we have seen, sometimes, getting labeled data in the target domain is challenging while labeled data exists in abundance in another source domain. This is when transductive transfer learning becomes useful. In some cases, the source and target feature spaces are different, and sometimes they are the same but the marginal probability distributions of the input data are different. The latter case of the transductive transfer learning is closely related to domain adaptation for knowledge transfer.In the subcase that scientists refer to as inductive transfer learning , however, it is the target and source tasks that differ from each other; in fact, sometimes even the source and target domains are not the same. In this case, some labeled data specific to the target domain remains necessary in order to induce an objective predictive model for the target domain.Finally, the unsupervised transfer learning setting allows (remarkably!) to reuse models even when the target task is different from but still related to the source task.click for larger image Figure 2: The different transfer learning strategies (Source: Figure Eight) Leave a Reply Cancel reply You must Register or Login to post a comment. This site uses Akismet to reduce spam. Learn how your comment data is processed. CameronRobertson says: “The algorithms behind these technologies is really astounding. It would be really amazing to put the machines to the test and see what they could really learn just by absorbing information around it. Of course we risk moving into a sentient AI phase of li What’s the best way to perform the transfer? https://www.infoworld.com/article/3155262/analytics/transfer-learning-jump-starts-new-ai-projects.html When is it appropriate to use it? They expect a better improvement rate (a better “learning curve”) of the source model than they would normally get training a target model using supervised learning on the target data set. July 26, 2018 at 11:21 am Log in to Reply It also offers an opportunity to save compute power, which is particularly important in the context of computationally heavy models.Why is Transfer Learning Important for IoT? A hairy problem when developing customized applications is known as the “cold start” problem. Cold start refers to a new user hitting a site where it is very difficult to build a personalized offering because his/her behavior or preferences are still unknown as no preliminary data has been collected. This is why your recommendations on an eCommerce website are usually disappointing until you start generating more data by browsing and shopping.IoT sees a similar issue. After all, it’s always hard to train a model for a new connected product due to the fact the new user’s preferences are unknown. Under these circumstances, a transfer learning approach can help because a warm restart typically requires much less data than would be necessary to recreate an entire model. Therefore, it offers an opportunity to ‘onboard’ new customers faster without the need to collect large amounts of data. In this scenario, the ‘base’ model would be generated using a combination of the data collected across a large number of customers which is more likely to be a fair representation of the expectations of the average user.Transfer learning is also critical to the successful deployment of IoT deep learning applications that require complex machine-generated information of such volume and velocity that it would be simply impossible to ever find enough human experts to label it and train new models from scratch in a reasonable amount of time. Think about it: if it takes days or even weeks to label data and train a particular model, by the time the model is ready for consumption, it is very likely to already be obsolete!No Magic Bullet Transfer learning truly opens the door to new applications, and is definitely one of the key elements for building smart IoT applications. However, some particular precautions need to be taken to ensure proper usage of this technology.With ever-growing concerns by consumers over the usage of their data, it is only natural to question the potential data security concerns related to IoT. As seen previously, transfer learning often means that a model built from an existing data set is used as the starting point to the development of a new solution. And this data isn’t necessarily open source data: it might sometimes be the aggregation of the data of a large group of people, or even (more rarely) the data set of a particular consumer. As a general rule, data scientists prefer to use aggregated data sets not only because it gives a more reliable worldview to start with when creating a model for a new customer, but also because the more data involved, the less likely it is that a specific individual might be directly identifiable.And there are also more technical challenges related to transfer learning. In fact, in some situations, knowledge should not be transferred at all, as brute-force transfer may hurt the performance of learning in the target domain, this situation is referred to as negative transfer . How to avoid negative transfer is an important open issue that is attracting more and more attention in the research community, including at Figure Eight.References: Jen McDade says: Continue Reading Previous The challenges of IoT security and how to harden the edgeNext ADLINK: smart gateway selected for Michelin-SoftBank tire control system CameronRobertson says: Jennifer Prendki is currently the VP of Machine Learning at Figure Eight, the Human-in-the-Loop AI category leader. She has spent most of her career creating a data-driven culture wherever she went, succeeding in sometimes highly skeptical environments. She is particularly skilled at building and scaling high-performance Machine Learning teams, and is known for enjoying a good challenge. Trained as a particle physicist, she likes to use her analytical mind not only when building complex models, but also as part of her leadership philosophy. She is pragmatic yet detail-oriented. Jennifer also takes great pleasure in addressing both technical and non-technical audiences at conferences and seminars, and is passionate about attracting more women to careers in STEM. Share this:TwitterFacebookLinkedInMoreRedditTumblrPinterestWhatsAppSkypePocketTelegram Tags: Advanced Technology, IoT October 6, 2018 at 2:10 am In short, data scientists expect an acceleration of their modeling productivity, a drop in the amount of data as well as the amount of compute required for their tasks, a more predictable development process and even some level of r isk mitigation since the approach or the data that is being transferred is already “true and tested”.Transfer Learning as a Ready-to-Go Model Technically, transfer learning is about encoding knowledge within the parameters of a model and making use of that knowledge to solve different problems. In practice, the transfer can occur in two different ways, starting with existing models built on preliminary data or with existing data alone. 4 thoughts on “Transfer learning for the IoT” Log in to Reply https://towardsdatascience.com/why-machine-learning-on-the-edge-92fac32105e6 A data scientist looks for an existing model that might appear as a good candidate to address his/her own needs. He/she might not even be aware of the data this model was originally trained on. In these scenarios, we talk about “pretrained models”, and the knowledge that is transferred is already encoded into the model itself, and the ML expert’s job is solely focused on fine-tuning the model using a small amount of data taken from the target domain through a process called warm restart . This is one of the most appealing ways to use the process, as it actually enables AI systems to learn from a small number of cases, which makes the training phase both fast and efficient. This process is revolutionary in the sense that it opens the door to collaborative artificial intelligence, when each developer or scientist can build on top of existing expertise, and focus solely on the added value that they can bring as an expert. “”sometimes it’s good to have a look back at our well-received articles from the recent past” — This is why I personally enjoy printing out some of my favourite articles. I’m not always meticulous about doing it because I’m pretty sure that there are a Log in to Reply “nThanks for sharing useful information about IoT. The Internet of Things (IoT) is an intense, transformative power and foundation for computerized organizations exploiting the meeting of the physical and advanced universes. Transfer learning will make Io RituGupta says: Log in to Reply August 16, 2018 at 6:26 am If you know anything about artificial intelligence and the internet of things, you most likely can’t help but feel a great deal of excitement at the idea of combining the extraordinary promises that these two fields have to offer. With the unparalleled advances in the field of computer hardware, some of the advanced theoretical knowledge that has been around for decades can be finally leveraged and put to use in real-life, practical applications. And with an ever-growing IoT market, getting high quality data has never been so easy, enabling the development of ever better, more performant models, and in particular of deep learning models.Wait, Before we Start: What is Deep Learning, Exactly? Deep learning is a family of artificial intelligence techniques responsible for countless technological advances in the most recent years. Even though deep learning (formally referred to as “deep neural networks”) has existed since the 50’s and has seen its key algorithmic breakthroughs in the 80’s, it is only recently that computer hardware advances have caught up to theories, giving computer scientists access to the enormous computational power and data storage that they needed to make neural networks reach their full potential. Deep learning algorithms are characterized by a unique multi-layer architecture in which knowledge is progressively “abstracted away” from the raw data they take as an input. For instance, when a deep neural net is trained to identify the pictures of a particular type of animal out of a pool of random images, it starts by first considering the images at the pixel level, then converts this knowledge into “higher-level” knowledge such as the edges of the object of interest, then identifies specific features (such as the paws, the tail, etc.) before eventually abstracting a general visualization of the object (or, to continue our example, of the animal) as a whole. The interesting point here is that the early layers that concern the lower-level knowledge could be reused with an entirely new type of images–such as satellite imagery for example–because edges remain edges regardless of what an image represents. Remarkably enough, no data scientist ever explicitly programed a computer to perform such a task. Instead, the learning algorithm is fed with hundreds of thousands of images until it figures out by itself (and learns by itself!) how to recognize the desired objects. Deep Learning on the Edge: Why it Matters In parallel to the hardware improvements that helped the advancement of deep learning research, some tremendous discoveries have been made that have allowed the rapid development of IoT devices. In particular, the preponderance of microcontroller units (MCU) are creating a unique opportunity to make AI applications–especially deep learning-enabled applications–accessible to the consumer at an unprecedented scale and speed.MCUs offer remarkable advantages for deployment of Deep Learning-based applications, as they reduce latency, conserve bandwidth and offer better guarantees in terms of privacy. When such AI applications are installed directly on an IoT device, that’s what I’d refer to as the deployment of AI applications on the edge. Choosing if an application is best deployed on server or cloud or on the edge of an IoT device comes down to making a decision regarding the trade-off between speed (through latency) and accuracy (as more complex and therefore larger models, can’t always be stored in the more limited memory of an IoT device).What is Transfer Learning? In the context of psychology, transfer learning is defined as the study of dependency of human learning or performance on preliminary experience. Humans are not taught how to perform every single possible task in order to master it. Rather, when they encounter new situations, they manage to solve problems in an ad-hoc manner by extrapolating old knowledge to new environments. For example, when a child learns to swing a tennis racket, he or she can easily transfer that skill set to ping pong or baseball. The same goes for conceptual understanding, like applying statistics or other mathematics to budgeting as an adult.By contrast to the way humans function, most machine learning algorithms, once implemented, tend to be specific to a particular data set or to a particular, discrete task. Machine learning researchers have been giving more and more attention recently to how to empower computers to reuse their acquired knowledge and re-apply it to new tasks and new domains, attempting to abstract and transfer the “smarts” extracted from the data across multiple applications, usually somewhat similar to each other.While Deep Learning can be used in an unsupervised context, generally it powers supervised learning. That means the algorithm learns from examples made of input-output pairs which it then uses to try to identify patterns and extract relationships, with the goal of predicting an outcome from new unseen data. Deep learning has application in countless areas of research, but suffers from one major drawback: these are extremely data-greedy algorithms require massive quantities of data in order to tune the thousands of parameters that come into play in a neural network architecture. This means that not only is a lot of data required to achieve good performance, but this data also needs to be labeled, which can be an expensive and time-consuming endeavor. Even worse, it is often not even possible to obtain quality data, or to label it accurately enough to be able to train a model from it.This is where transfer learning can really help. It allows for a model developed for a specific task to be reused as the starting point on another one. It is an exciting area for machine learning scientists because it mimics the way humans generalize knowledge from one specialized task to another. In fact, it is a key strategy when it comes to reducing the required size of datasets in order for deep neural networks to even become a viable option.Remember the example we surfaced earlier about how a model can learn to identify object edges in a deep learning network? That’s the exact sort of knowledge that can be transferred to another computer vision task. In this way, a model that learns to see regular objects like cats and dogs can transfer that understanding to more complex tasks, like identifying nuclei in cancer imagery.click for larger image Figure 1: The different learning approaches when building AI applications (Source: Figure Eight) Transfer Learning Strategies Transfer learning certainly comes across as an elegant and natural way to solve the dilemma of information poverty in the age of Big Data. However, some important tactical questions regarding transfer learning need to be answered prior to being able to use it in practice: “We have smaller instances of something that is already combining these 2 technologies in Google Home and Alexa and things like that. It’s a good introduction to just how much more sophisticated such collaborations can become in the future. I mean I person They expect a stronger starting point, meaning that the initial predictive power (prior to refining the model) on the source model is higher than it otherwise would be. A data scientist realizes that the data available to train a supervised model is insufficient in quantity and/or quality, or that labeling the data is too tedious or expensive, and decides to build a model from scratch by leveraging another existing labeled data set originally developed in the context of another task or another domain. This allows him/her to make use of existing labels in order to save money or gather labels more easily. In this case, the data scientist starts with a regular supervised learning approach solving a different problem (a different task), or the problem at hand using another data set. He/she then finishes by fine-tuning the model to best adapt it to his/her problem.
In a repeat of last season, Haryana Hammers will take on Punjab Royals in the Pro Wrestling League (PWL) final following their 6-3 win over Delhi Sultans in Greater Noida on Wednesday.Local boy Rajneesh clinched the men’s 65kg category bout against Andrey Kviatkovski of Ukraine 9-3 to give the last year’s finalists an unbeatable 5-2 lead over Delhi Sultans with two bouts remaining.Rajneesh hit the final nail in Delhi’s coffin after Ravi Kumar, Kiran, Anastasia Nichita and Aleksander Khotsianivski won four off the six bouts played.Ali Shabanov added one more win in Haryana’s kitty when he got the better of Sanjeet Kundu 5-0 in the men’s 86kg category after Sakshi Malik pulled one back for the Delhi Sultans beating Tatyana Omelchenko 6-2 in women’s 62kg category.Earlier, World U23 Championships silver medallist Ravi Kumar put the Haryana Hammers ahead winning the men’s 57kg bout against Pankaj 7-1 which included a beautiful four-point takedown in the first round.The 2018 CWG bronze medallist Kiran, showing tactical superiority, made it 2-0 in Haryana’s favour defeating the 2018 European U23 champion Anastasia Shustova of Ukraine 6-1 in the women’s 76kg category.The Russian national champion Khetik Tsabolov made it 1-2 for Delhi Sultans defeating Haryana Hammers’ 2017 Commonwealth Championships silver medallist Praveen Rana 9-0 in the men’s 74kg category.The world junior champion Anastasia Nichita of Moldova gave Haryana Hammers a healthy 3-1 lead, getting the better of Delhi Sultans’ Katherina Zhydachivska 11-2 in what turned out to be a one-sided encounter in the women’s 57kg category.advertisementAleksander Khotsianivski took the Hammers one more step closer to winning the tie by clinching the men’s 125kg superheavyweight bout against Satender Malik of Delhi on technical superiority. Aleksander won the bout 16-0 in 51 seconds.National champion Pinki of Delhi delayed the inevitable by winning the women’s 53kg bout against Haryana’s 2017 Commonwealth Championships gold medallist Seema 3-1.
New Zealand cannot afford to squander even half-chances against a team of India’s calibre in Thursday’s World Cup match at Trent Bridge, Kiwi fast bowler Lockie Ferguson said.The right-arm paceman will be spending much of his 28th birthday trying to tame a formidable Indian batting order headlined by Virat Kohli and Rohit Sharma, the top two batsmen in ODI rankings.”They’re world-class players, you’re not going to blow them out of the water, but if you can build up enough pressure against them and then create a half-chance, that could be the wicket and you can then build from there,” Ferguson told reporters on Tuesday.”Obviously, they’re playing some great cricket and they’re one of the top teams in the competition but we’re definitely looking forward to the opportunity of playing them in England and we haven’t played them for a while in England.”India have adopted a new approach to the first powerplay overs which was evident in their clinical victory against Australia on Sunday.Kohli’s men collected only 41 runs in the first 10 overs, without losing a wicket, but went on to plunder 116 during the last 10 en route to a commanding 352-5 which was beyond Australia’s reach.”They showed us that they’re probably a lot more patient in a way,” Ferguson said.”And although we’re looking to take wickets, sometimes we got a little bit expensive. I think taking wickets up front is the key to (beating) India but, if not, creating pressure and building dots balls.”The threat of rain looms large over the contest between the tournament’s only two unbeaten teams. A record three matches have already been either abandoned or called off in the weather-hit tournament and Ferguson hoped Trent Bridge would not fall victim to another downpour.advertisement”It’s the World Cup. We’re playing against India in the World Cup and it’s an opportunity to get two points and we don’t want to get rained-out games,” he said.”I don’t think any players do but if that happens then so be it. We can’t control that but we’re looking forward to playing India and getting some confidence against them.”Also Read | Shikhar Dhawan joins MS Dhoni, Hardik Pandya and KL Rahul to watch Bharat in NottinghamAlso Read | Rishabh Pant to join Team India as cover for injured Shikhar DhawanAlso Read | Shikhar Dhawan’s absence will put more pressure on Rohit Sharma, says Harbhajan SinghAlso Read | Shikhar Dhawan under observation, no replacement yet
I’m not besmirching what Childs did. He was great, and good for him because he’s been with the program through some down times. It was the best college football game he has ever played. That’s awesome.And you guys know my feelings on Hill. I think he’s great. I think he’s going to be great and is already great. I think he’s the next Joe Randle. He also had some really good runs on Saturday.But none of this means the run game woes are no more.In fact, let’s look at those here. I went back and looked up the total yards-per-carry totals for OSU in its three non-conference games dating back to 2010 (I only included the first three non-con games of 2010 to try and accurately match it up with the other years). OK, here we go.#okstate yards per carry in non-conference play … 2010: 5.32011: 4.72012: 7.02013: 4.82014: 4.22015: 4.92016: 3.1?— Pistols Firing (@pistolsguys) September 19, 2016 Rennie Childs ran for 101 yards and four TDs on 10 carries on Saturday against Pitt. That’s terrific. Really good. The first time an OSU RB has run for 100+ against a Power 5 opponent since, gulp, the 2015 Cactus Bowl. That was two seasons ago.But is the run game even going in the right direction? Mike Gundy thinks so.“Well, as I mentioned earlier, it’s not good enough, but we’re just one guy away,” said Gundy. “One play, it’s the right guard. The next play everybody is blocked, it’s the left guard. Another play, everybody is blocked, and it’s the right tackle.“We’re just one guy away on a lot of plays from being better in the running game. Over the last two years, we were three guys at a time away up front. We’re making improvements. Rennie (Childs) played the best game he’s ever played. He showed speed, he showed some allusiveness. Chris Carson is hurt, so we don’t have him for a while. I liked the way Justice ran. He was close to breaking a couple. He did a nice job for a true freshman.” Oh my goodness.OSU has been a full yard per carry worse in non-conference play this season than in any season since Dana Holgorsen took over the offense in 2010. You can’t blame the schedule either. Yeah, Pitt is pretty good, but OSU played Florida State in 2014. It played Mississippi State in 2013. It played Arizona in 2012. These teams were all as good or better than Pitt.OSU also currently ranks No. 113 in the nation in sacks allowed (OU is No. 117 by the way). And this is part of the problem. If you take out Rudolph’s “runs” OSU jumps up to 4.5 yards per carry.But I didn’t take out the QB “runs” in any of the other seasons either. Sacks were fair game there as they always are (although the QBs in those seasons weren’t sacked as much). And it’s not like the sacks don’t count in the game! The refs don’t say, “well the QB had it and he wasn’t trying to run so it doesn’t count.” It’s just that you would categorize the issues under “offensive line” instead of “run game.”So all of this is a problem. One we haven’t really discussed that much because, you know, the run game seems better for the most part. Gundy says they’re one guy away on most plays, and maybe that’s true. But in terms of pure numbers, things are somehow getting worse and not better.UPDATE (2:33 p.m.): I went back and took out sack numbers for the last three years. Still not great. And my point still remains. The running game hasn’t gotten better statistically.2014: 4.8 YPC2015: 5.4 YPC2016: 4.2 YPCIf you’re looking for the comments section, it has moved to our forum, The Chamber. You can go there to comment and holler about these articles, specifically in these threads. You can register for a free account right here and will need one to comment.If you’re wondering why we decided to do this, we wrote about that here. Thank you and cheers!
So we’ve reached the final day of the X-Blades 18 Years and Under National Championships and it’s going to be another day of dynamic competition. There are only four teams left in the hunt for the Boys’ Championship.Reigning champions Sydney Mets will play Cobras in the first semi-final while QSST will face off against Central Queensland on the other side of the draw. Mets will take a mental advantage into their match, having already disposed of Cobras 8-4 on Thursday, but the Cobras boys led by the impressive Rhys Jacks and Oscar Sanft will be steeled for a battle to the end. QSST will be hunting for their first finals appearance in a few years when they go toe to toe with the form team of the 2007 tournament Central Queensland.In the Girls’ Championship race gutsy performances by Queensland franchises Sharks and Cobras earned them a place in the final four. Cobras, led by Australian 20 Years gun Sam Hopkin, upset a powerhouse Suns team 4-3 while Sharks easily disposed of the higher-placed NSWCCC 8-2 who will be disappointed with their meek surrender after promising so much in their narrow loss to QSST yesterday afternoon.Sharkies will battle QSST and Cobras will face NSWCHS in the semis. The Boys’ Shield semi-finals will see more hotly contested matches. Northern Eagles will be disappointed at missing out on the Championship finals. The home-town heroes will be desperate to progress through to the Shield grand final but they will first have to overcome Sunshine Coast. Sharks and Suns are the other two teams left in the battle to claim the boys’ Shield. In the girls’ division Central Queensland earned itself a position in the final four with a 6-5 drop-off win over Northern Eagles last night. CQ will meet Sunshine Coast for a place in the Shield grand final. Whoever makes it through will come up against the winner of NSWCIS and Hunter Hornets.The boys’ Plate finals will also be contested tomorrow. Rebels and Tasmania will square off for one grand final berth while Scorpions and Northern Territory will clash for the other. It is an all-foreign battle for the girls’ Plate. ACT and TouchWest will play in the first semi while Tasmania and Northern Territory will meet in the second. The girls’ grand final is scheduled for 2.15pm. The boys will play off at 3.15pm. Every team competing in the X-Blades National 18 Years and Under Championships still has something to play for. Who will finish champions? We will know by the end of today. 8.00amGirls’ Shield SF1 Sunshine Coast v CQGirls’ Shield SF2 NSWCIS v HornetsGirls’ Plate SF1 ACT v TouchWestGirls’ Plate SF2 Tasmania v Northern TerritoryBoys’ Plate SF2 Rebels v Tasmania9.00amGirls’ Champ SF1 QSST v SharksGirls’ Champ SF2 NSWCHS v CobrasBoys’ Shield SF1 Northern Eagles v Sunshine CoastBoys’ Shield SF2 Sharks v SunsBoys’ Plate SF1 Scorpions v Northern Territory10.00amBoys’ Champ SF1 Mets v CobrasBoys’ Champ SF2 QSST v CQBoys’ P/O 7/8 NSWCCC v HornetsBoys’ P/O 13/14 TouchWest v NSWCISGirls’ P/O 13/14 Rebels v Eagles11.00amBoys’ P/O 5/6 ACT v NSWCHSGirls’ P/O 5/6 Suns v NSWCCCGirls’ P/O 7/8 Mets v ScorpionsGirls’ P/O 11/12 L/S/SF1 v L/S/SF2Girls’ P/O 17/18 L/P/SF1 v L/P/SF2 12.00pmGirls’ Shield Final W/S/SF1 v W/S/SF2 (Determines 9th and 10th)Girls’ Plate Final W/P/SF1 v W/P/SF2 (Determines 15th and 16th)Boys’ P/O 11/12 L/S/SF1 v L/S/SF2Boys’ P/O 17/18 L/P/SF1 v L/P/SF2 1.00pmBoys’ P/O 3/4 L SF1 v L SF2Girls’ P/O 3/4 L SF1 v L SF2Boys’ Shield Final W/S/SF1 v W/S/SF2 (Determines 9th and 10th)Boys’ Plate Final W/P/SF1 v W/P/SF2 (Determines 15th and 16th) 2.15pmGirls’ Final W/SF1 v W/SF23.15pmBoys’ Final W/SF1 v W/SF2
Editors’ Recommendations Much has been said and written about rooftop bars around the country. Who doesn’t love having a drink outdoors while the city throngs teem far below? But the concept isn’t limited to just swanky hotels or apartment complexes. Airports can be stuffy and even the nicest first-class lounge can feel a bit sterile at times. Enter Delta Air Lines’ new outdoor Sky Deck at Terminal 4 at JFK International Airport. It’s anything but sterile.The 2,000-square-foot terrace is part of the new Sky Club and was developed in conjunction with Architectural Digest and interior designer Thom Filicia. Guests get unimpeded views of the runway, a first at JFK. The space is outfitted with comfortable JANUS et cie MOOD chairs and couches, perfect for those long layovers when all you want is to sink into a seat. Safavieh rugs dot the floors and Antonangeli, FLOs, Holly Hunt and Hinkley lights cast a ethereal glow. A small bar and potted plants complete the space.“Amid the hustle and bustle of the airport experience, we seek to create a refuge, which represents Delta’s dedication to elevated design and pays homage to an elegant lifestyle of travel and the spirit of flight,” Filicia said.The Sky Deck opened early this summer and is part of the new Delta Sky Club at the revamped Terminal 4. At 24,000-square-feet, the Sky Club is the largest in the Delta fleet. Guests can order food at both the Sky Deck and the interior Sky Club. Sushi, cheese, charcuterie and dessert are all available as is premium wine selected by Master Sommelier Andrea Robinson. In June, Delta launched its sister Sky Deck at Hartsfield-Jackson Atlanta International Airport. Smoking is prohibited at both lounges. Day passes can be purchased for $50. The Best Podcasts to Listen to During Your Workout Mountain Hardwear Is Bringing Augmented Reality to the Great Outdoors Stirred or Shaken? How to Make a Perfect Vodka Martini Redefining Rustic, Rancho Buena Vida is the Classic Texas Ranch Made Over The Best Steakhouse Whiskey Collections in the United States
Construction of Nova Scotia’s new liquefied natural gas (LNG)terminal and re-gasification plant at Bear Head is set to beginthis summer. Nova Scotia Business Inc (NSBI) has announced completion of thesale of an industrial site at Bear Head, Richmond Co., for $4.62million to the Bear Head LNG Corporation, which will build andoperate a $600-million LNG terminal and re-gasification plant onthe site. Stephen Lund, president and CEO of NSBI, made the announcementtoday, May 2. The property includes 179 acres of land and a67-acre water lot. “The completion of the land sale represents a major milestone inthe life of the project,” said Mr. Lund. “Construction of thefacility will begin this summer, and we look forward to the firstflow of gas into the pipeline from the Bear Head LNG plant in2008.” Acting as the project manager for the Bear Head LNG opportunityon behalf of the province, NSBI completed the transaction as partof its energy-sector strategy. In order to maximize the potentialeconomic gains from the LNG opportunity, NSBI is also activelypursuing new, promising gas-related commercial opportunities. The Bear Head LNG terminal and re-gasification plant will ensurea long-term gas supply to the province and the northeast UnitedStates. The operation, taken together with Nova Scotia’s currentoffshore gas resource base, will result in the expansion ofexisting pipeline infrastructure into the New England market andenhance the province’s position as a key energy supply region inNorth America. “This project in the Strait of Canso region will have a majorimpact on the economic future of the province and the energysupply chain in North America,” said Mr. Lund. “Communityleadership, a world class site, and strong commercial partnerswith a solid business plan combined to make this massive projecta reality.” Following an August 2004 project announcement, the industrialsite and access road were prepared for the beginning of plantconstruction, which will start in the summer of 2005. Theterminal and re-gasification plant will be operational by mid2008. Re-gasification is the process by which liquefied naturalgas is converted back to a gaseous state prior to transmissionthrough a pipeline. The three-year construction phase of the project will create anaverage of 600 jobs, with 35 to 40 permanent positions whenoperations begin. Nova Scotia Business Inc. is the business development agency forNova Scotia, led by a private-sector board of directors. It isthe investment attraction arm of the province and helps companiesin Nova Scotia meet their growth potential through businessadvisory, export development and financing services.
Premier Rodney MacDonald offered congratulations today, Aug, 29, to 28-year-old Dwight D’Eon, who ended his Canadian Idol journey last night. Mr. D’Eon, a lobster fisherman from West Pubnico, was a top four finalist on Canadian Idol, before his elimination from the popular TV show Tuesday night. “All Nova Scotians are proud of Dwight, his superb talent and his achievements on Canadian Idol,” said Premier MacDonald. “In our eyes, ears, minds and hearts, he is undoubtedly number one.” Mr. D’Eon, was recognized for his exceptional guitar skills during the contest, which started last spring with thousands of amateur musicians from across Canada trying to become a Canadian Idol. DOR BROADCAST USE: The province today (August 28th) congratulated Canadian Idol top four finalist Dwight D’Eon on behalf of all Nova Scotians. The 28-year-old lobster fisherman from West Pubnico was eliminated from the popular television show Tuesday night. Premier Rodney MacDonald says all Nova Scotians are proud of Dwight and that he is number one in our eyes, ears, minds and hearts. -30-
REGINA — Conservative Leader Andrew Scheer welcomed 2019 with a warning that if Canadians re-elect Justin Trudeau this year, the federal carbon tax that’s going to take effect will only climb.“Canadians know what Justin Trudeau is going to do. Now that his carbon tax is here, it’s only going to go up. And if he gets re-elected in 10 months, it will go up even more,” Scheer said during a New Year’s Day news conference in a Giant Tiger store in Regina.“This time next year I plan on being able to tell Canadians that Justin Trudeau’s carbon tax is a thing of the past.”The federal government’s new carbon pricing system comes into effect in 2019 in provinces that don’t have carbon pricing mechanisms of their own. The carbon price outlined by Ottawa starts at a minimum of $20 a tonne and rises $10 annually until 2022.But Scheer told reporters that government officials are saying the tax would need to rise to $100 per tonne for it to be effective at reducing carbon emissions, and he says the federal environment department is planning for a carbon tax of $300 per tonne.“So we know Justin Trudeau will raise the carbon tax higher. His experts are telling him to. His own government departments are telling him to,” Scheer said.“At that price home heating bills will rise by more than $1,000 a year and gas prices would go up by more than 60 cents a litre.”When asked by a reporter about his own plan to fight climate change, Scheer responded that the Conservatives’ plan will help reduce global emissions by capitalizing on Canada’s clean technology and cleaner energy, which he said will also keep manufacturing jobs in Canada instead of moving to countries without those things.Canada’s former parliamentary budget officer predicted in a report in April that the federal government’s carbon tax will cut economic growth by 0.5 per cent or $10 billion dollars when it’s fully implemented in 2022, and would generate significant revenues.However, Jean-Denis Frechette’s report noted the impact on the economy will depend on how those revenues are used.Trudeau says Ottawa will return 90 per cent of the money it collects from a carbon tax to Canadians.Saskatchewan is asking its Court of Appeal to rule on whether the carbon tax is unconstitutional and has argued its climate change plan is enough to reduce emissions and a carbon tax would hurt the Saskatchewan economy.Scheer, who represents Regina—Qu’Appelle in Parliament, said Saskatchewan’s fight against the tax gets easier as more provinces elect governments that also oppose it, like Ontario and New Brunswick did in 2018.“Premier (Scott) Moe and Premier (Brad) Wall before him had a bit of a lonely battle but now they’ve got reinforcements. And my message to Canadians is come 2019, the battle will be won,” Scheer said.The Canadian Press
Tangier – Abdallah Ouzad, the Amazigh actor and theater director was found dead under suspicious circumstances Sunday morning in his room at the Cesar Hotel in Tangier, where he was attending the Tangier Film Festival.The local police are investigating to determine the cause of death once an autopsy of the body is performed.According to Al-Akhbar daily,the 54 year old actor was found dead in room number 101 by the hotel’s housekeeping staff, but the cause of death is unknown. According to the same source, the initial investigations by the police seem to suggest that Ouzad died of heart failure, and that he suffered from asthma.Ouzad leaves behind a huge artistic legacy. Starting out as an amateur actor in Sale, playing in the successful movie “Horses of God” and directing the Amazigh theater “IKBOUREN,” Abdallah Ouzad joins the list of the great Moroccan artists who have left the cinematic scene.Ouzad was born in Sale in 1960, and was the husband of the Moroccan painter Sharifa EL-HAMRI. He is considered to be the founder and the godfather of the Amazigh theater in Morocco. Edited by Elisabeth Myers © Morocco World News. All Rights Reserved. This material may not be published, rewritten or redistributed
27 October 2009At least 10 people have been killed in a new round of inter-tribal clashes in Darfur, the joint African Union-United Nations peacekeeping force to the war-scarred region of western Sudan reported today. The mission, known as UNAMID, said 10 members of the Birgid tribe died after clashing yesterday with Zaghawa tribesmen near Shangil Tobaya, a town about 70 kilometres from El Fasher, the capital of North Darfur state.Twelve other Birgids were injured as a result of the clashes and they are being treated either at UNAMID medical facilities or a clinic run by the non-governmental organization (NGO) known as Médecins Sans Frontières (MSF).Yesterday’s clashes occurred less than a week after two people were killed when Birgids and Zaghawas engaged in fighting, also near Shangil Tobaya.UNAMID has attributed the recent fighting to disputes over access to water, a particularly precious commodity in Darfur, an arid and landlocked region.The mission is sending an assessment team to the area tomorrow comprising military officers, police officers and humanitarian, human rights, security, civil affairs and protection officials.UNAMID has been in place across Darfur since the start of last year to try to quell the fighting and humanitarian suffering that has engulfed the region since 2003. At least 300,000 people are estimated to have died from the conflict and another 2.7 million people remain displaced from their homes.
Tensions escalated between the two neighbouring South-East Asian nations in July 2008 following the build-up of military forces near the Preah Vihear Temple in Cambodia. The temple, which dates back to the 11th century, was inscribed on the World Heritage List of the UN Educational, Scientific and Cultural Organization (UNESCO) earlier that month. The row turned deadly in October of that year when two people reportedly died during an exchange of fire between Thai and Cambodian forces near the site. “The Secretary-General hopes that Cambodia and Thailand will resolve the dispute along their border amicably through dialogue,” UN spokesperson Farhan Haq told reporters in New York. “He stands ready to help the parties,” Mr. Haq added. 20 August 2010Secretary-General Ban Ki-moon today offered to help Cambodia and Thailand resolve their differences over a disputed temple area along their common border.
Families opposed to the erection of large-scale wind farms near their homes failed to prove the projects would cause any serious harm to their health, an Ontario government lawyer said Tuesday.In his opening comments, Matthew Horner told a Divisional Court panel that a review tribunal was correct to reject objections to the turbines based on health concerns.“There’s no indication that the tribunal made a palpable and overriding error,” Mr. Horner said late on Day 2 of the three-day hearing.He also said the tribunal was right to reject the residents’ “novel argument” that the approvals process violates the constitution.Four families are asking the appellate court to throw out decisions by the Environment Review Tribunal that upheld approvals of three large-scale wind-energy projects.They also want the approvals process declared unconstitutional on the grounds that the law precludes them from arguing turbines might cause them harm.In her submissions, another government lawyer, Danielle Meuleman, rejected suggestions the tribunal hearings had been unfair to the families.Among other things, the families’ lawyer had told the court the tribunal was wrong in failing to grant a requested adjournment to allow time to get expert evidence together.Ms. Meuleman countered that the tribunal took into account “all the procedural rights” of everyone involved in deciding against stopping the process.She also dismissed suggestions the tribunal had “failed” to allow the families to enter a Health Canada study on the effects of wind turbines as evidence.A summary of the study, released Nov. 6, found no direct link between turbines and the health of nearby residents but did find a link to their levels of “annoyance” which could have adverse health effects.“They never asked the tribunal to stop the clock and wait for the results of the Health Canada study,” Meuleman said. In addition, she said, only a summary of the study has been released and no one knows what exactly it means.“We’re all speculating,” Ms. Meuleman said.Earlier in the day, the lawyer for the families wrapped up his submissions by asking the court to order the Environmental Review Tribunal to hold new approval-review hearings on the projects.“Send it back with constitutional relief,” Julian Falconer told the justices.The constitutional relief, Mr. Falconer said, would involve “reading down” the relevant legislation, the lawyer said.Justice David Brown asked what the altered legislation might look like and Falconer said it should create a “reasonable prospect of harm” as a ground to challenge wind-energy projects.As it now stands, opponents have to prove they have suffered actual harm before they can stop a project, the lawyer said.“I understand it’s not much of a test if you first have to get sick in order to prove it,” Associate Chief Justice Frank Marrocco observed.The case is slated to wrap up Wednesday, with the government and project proponents making further submissions.Canadian Press
SMITHS FALLS, Ont. — Canopy Growth Corp. says it generated $97.7 million of revenue in the three months ended Dec. 31, including its first sales of non-medical cannabis products, up from $21.7 million in the previous year’s fiscal third quarter.The Ontario-based company — one of Canada’s largest cannabis producers — recorded a loss from operations for the quarter, but turned a net profit due to gains on the fair value of its assets and liabilities.Loss from operations was $157.2 million, compared with a loss of $26 million a year earlier. Companies backed by major cannabis producers are inking deals with Ontario retail lottery winners Aurora Cannabis CEO: ‘I lose sleep over our ability to supply this global cannabis market’ The cannabis industry could be worth billions less than Canada thought Net income attributable to shareholders, including net gains on the fair value of its assets, was $67.6 million or 22 cents per basic share, up from $1.6 million or one cent per basic share a year earlier.Adjusted EBITDA (earnings before taxes, depreciation and amortization) was negative $75.1 million, compared with a year earlier adjusted loss of $5.68 million.Canopy also reported that if all of its convertible debt had been converted to equity at the beginning of the quarter, it would have had a net loss of 38 cents per diluted share.The company said Tim Saunders has decided to retire as chief financial officer later this year after assisting with the transition to a new CFO, but will remain on the board of directors as a strategic adviser.
Click to share on Facebook (Opens in new window)Click to share on Twitter (Opens in new window) SMMT will be holding a webinar on Thursday 18 April to explain and discuss government’s introduction to tax credit for large company R&D expenditure on 1 April 2013.On the panel:Yung Tran: Head of Member Services and Business Improvement, SMMTRobert Baker: Chief Economist, SMMTDiarmuid MacDougall: R&D Partner, PwCThe webinar will be co-presented by Robert Baker and Diarmuid MacDougall. Robert will provide an overview of automotive and motoring related items from the 2013 government Budget. Diarmuid MacDougall, R&D Partner at PwC, has been heavily involved in the design of the new R&D credit.On 1 April the government will introduce a new R&D credit, which can be treated like a grant and set against the cost of investment by companies in their research or development. This comes after a huge number of submissions by businesses and industry bodies, particularly SMMT and EEF, calling for the change.While the government already provides an R&D benefit, it is accounted for by businesses as a tax which means it doesn’t reach the R&D budget holders who are responsible for R&D investments. This lack of visibility means it has much less impact on investment decisions. Worse still, it only provides cash savings if a company owes HMRC corporation tax on its profits – which may not be the case if the business has been investing heavily in R&D technologies, which many are, or if the business incurred losses in the downturn, as many did.The webinar will cover:Activities that will qualify for R&D credits.The expenditure on which the credit is given.How the existing R&D relief and the new grant like R&D credit work and issues to address when electing into the new scheme.Methods for making R&D claims which minimise the effort involved.The webinar will take place on Thursday 18 April, 10:30-11:30. Presentations will be followed by a Q&A session.To register, click here, fill in your details and joining instructions will be e-mailed to you directly
Senior safety C.J. Barnett jumps over a Michigan State player during the Big Ten Championship in Indianapolis Dec. 7. OSU lost, 34-24. Credit: Shelby Lum / Photo editorFORT LAUDERDALE, Fla. — Ohio State’s focus might not be completely on the game the day before the battle for the Orange Bowl. Split between hoping to snag a win and battling health, suspensions, injuries and new positions, players’ and coaches’ focuses are pulled thin.“We’ve got to get some guys healthy. It’s pretty much through our team now. We had a week-long battle with some virus stuff,” coach Urban Meyer said in a press conference Thursday. “Our Jeff Heuerman seems to be the last casualty that we’re working through.”As the 24-hour bug makes its way out of the picture, casualties in the lineup still persist. A three-game suspension has sophomore defensive lineman Noah Spence sitting out, and redshirt-junior cornerback Bradley Roby is still pushing through a knee injury sustained during the Big Ten Championship.Two starters out means Meyer has had to shift players to unfamiliar territory.“(Redshirt-freshman defensive back) Tyvis Powell and (freshman defensive back) Vonn Bell will be starting at positions they’ve never started before … We’ve made some adjustments to scheme,” Meyer said.As the numbers across the field change, Meyer has had to place faith in players who haven’t seen quite as much playing time. Sophomore defensive end Jamal Marcus has yet to start a game in the 2013 season and has recorded 15 tackles in contrast with Spence’s 52 tackles.“Jamal Marcus is going to be a disruptive guy. He’s one of the more talented guys on our team. I’m anxious to watch him play,” Meyer said. “He’s a quick twitch guy. This is his kind of game.”With Marcus geared up to take Spence’s place, the uncertainty of Roby’s injury still lingers. The cornerback trails only junior linebacker Ryan Shazier and redshirt-senior safety C.J. Barnett with 69 tackles. Only 10 tackles stand between Roby and the No. 2 position on the team.Defensive coordinator and linebackers coach Luke Fickell said staff would be taking the injury day-by-day, yet there still has been no word on Roby’s health status.Roby’s injury is another reminder of the loss to Michigan State Dec. 7, and that defeat is still on the minds of many.“You get back, and you see the players you care about and see the pain on their face. We had a real emotional meeting. I don’t know if emotional is the right word, but just like you would with any type of family members going through a hard time. From that point forward, they’ve been fine,” Meyer said.The coaches and players are attempting to shake off the loss (a feeling unknown to many newcomers on the team) and silence criticism of the formerly long undefeated Buckeyes.Kickoff between the No. 7 Buckeyes (12-1, 8-1) and the No. 12 Clemson Tigers (10-2) is set for 8:30 p.m. Friday at Sun Life Stadium.This article has been revised to reflect the following correction:Correction: January 2, 2014An earlier version of this article stated that the Orange Bowl would take place Thursday Jan. 2, when it in fact is scheduled for Friday, Jan. 3.
The head of MI6 will on Monday highlight the urgent need for a new era of spying in which artificial intelligence and robotics are deployed to combat rogue states hellbent on “perpetual confrontation” with the UK.In a rare public speech – only his second in four years in the job – Alex Younger, the Chief of MI6, will say that Britain must enter an age of “fourth generation espionage” to keep the country safe.The MI6 boss – known as “C” – will also emphasise the importance of “strengthening” Britain’s security ties with European allies ahead of Brexit, pointing out that “multiple” Islamic State-inspired attacks on the Continent have been disrupted thanks to the co-operation of intelligence agencies.The speech to students at St Andrew’s University, where Mr Younger studied, will also warn of the danger of “adversaries” who are “willing to take advantage” of huge leaps in cyber technology to launch attacks on Britain “in ways that fall short of traditional warfare”.Mr Younger will single out Russia for its “malign behaviour”. The Kremlin had ordered the assassination of Sergei Skripal, a former colonel in Russian military intelligence who spied for MI6, using nerve agent smeared on his front door handle in Salisbury. MI6 chief Alex Younger Want the best of The Telegraph direct to your email and WhatsApp? Sign up to our free twice-daily Front Page newsletter and new audio briefings. Mr Younger will say: “The era of the fourth industrial revolution calls for a fourth generation espionage: fusing our traditional human skills with accelerated innovation, new partnerships and a mindset that mobilises diversity and empowers the young.”The fourth industrial revolution, that follows the digital revolution of the internet age, is expected to see huge breakthroughs in the widespread use of artificial intelligence, robotics, nanotechnology and quantum computing.Mr Younger will say the combined response to the Salisbury attack in March had “exposed the perpetrators” and led to dismantling of Russian spy networks in a host of countries through the expulsion of diplomats. Mr Younger will “urge Russia or any other state intent on subverting our way of life not to underestimate our determination and our capabilities, or those of our allies”. MI6 will also need to recruit a new generation of younger spies, from diverse backgrounds, going forward. Aimed at students, he will say: “I want to speak to young people who have never seen themselves in MI6… it doesn’t matter where you are from. If you want to make a difference and you think you might have what it takes, then the chances are that you do have what it takes, and we hope you will step forward”. The public comments will reinforce concerns at the highest level that the rules-based international order is being flouted by the likes of Vladimir Putin, Saudia Arabia’s crown prince Mohammed bin Salman and President Donald Trump. Sergei Skripal with his daughter Yulia. They were both poisoned in the nerve agent attack in Salisbury
A supernova was discovered yesterday closer to Earth than any other of its kind. Astronomers think they found the supernova within hours of it exploding, and discovering the supernova so early and so close to Earth has had the astronomical community abuzz with excitement.So what exactly is a supernova? It’s is a stellar explosion that causes a burst of radiation that can often outshine an entire galaxy before fading from view over a couple of weeks or months. During this relatively short time, a supernova can radiate as much energy as the Sun is expected to emit over its entire life span. (That’s a lot of energy.) The explosion ejects much or all of the star’s materials at a speed of up to 30,000 km/s.Astronomers at Lawrence Berkeley National Laboratory at UC Berkeley made the discovery and think it will be one of the most-studied supernovae in history. Joshua Bloom, assistant professor of astronomy, called it “the supernova of a generation.”The exploding star, named the PTF 11kly, happened in the Pinwheel Galaxy in the Big Dipper. The Palomar Transient Factory survey (PTF) discovered the exploding star soon after it had burst. Scientists in the PTF have discovered over 1,000 supernovae since it started operating in 2008, but they think this one may be their most significant discovery yet.The PTF survey scans the skies with a robotic telescope mounted on the 48-inch Samuel Oschin Telescope at Palomar Observatory in Southern California. The PTF survey sends the observations more than 400 miles to the National Energy Research Scientific Computing Center (NERSC). Here, computers scan through the data and identify events that look like they need to be followed up on. The automated system sent out coordinates to telescopes around the world for observations within hours of the explosion.The supernova was categorized as a Type Ia, which is the kind used to measure the expansion of the Universe. The last Type Ia supernovae were back in 1972, 1937, and 1572. Weidong Li, senior scientist at UC Berkeley said they still don’t know what causes supernovas, but the PTF 11kly may help them figure that out. The team sent an urgent request to NASA’s Hubble Space Telescope to begin studying the supernova’s chemistry and physics starting this weekend.The PTF 11kly is getting brighter by the minute. It’s already 20 times brighter than it was yesterday. The supernova is 21 million light-years from Earth, which is a relatively small distance by astronomical standards.As the supernova gets brighter, it might even be noticeable with binoculars in 10 days from now. The best time to see the exploding star is just after evening twilight in about a week. You’ll need dark skies, but it’s said to be brighter than any other supernova of its kind in the last 30 years, so a pair of good binoculars should do the trick.via Berkeley LabAlso, with Hurricane Irene on everyone’s mind, doesn’t the above picture of the supernova look at all like Hurricane Irene?Photo via NASA
LA CENTER — It’s not uncommon to see a fridge in a classroom at La Center Middle School and La Center Elementary School.Teachers put apples, carrot sticks and peanut butter and crackers in them to hand out to students as snacks on occasion. Because the K-8 campus is bursting at the seams, the school staggers lunch periods throughout the day, meaning some students are eating lunch as early as 10:30 a.m. while others don’t get into the cafeteria until shortly before 1 p.m.The K-8 campus currently houses 1,736 students, 375 of whom are considered “unhoused,” meaning they have class in one of nine portable classrooms at the school, according to Superintendent Dave Holmes. There are also two portable classrooms at La Center High School.“We have no empty classrooms in the entire district,” Holmes said. “I have bids on my desk for bringing in another portable next year (at the K-8 campus), and we don’t have room for it. If we get another portable, we’ll probably have to put it on the football field.”The district will turn to the community to help with overcrowding on Feb. 13, when the district will ask residents to vote in favor of a $48,093,000 bond measure. Most of the money will go toward the construction of a new 81,375-square-foot middle school. A little more than $1 million of the bond money would go toward renovating the K-8 building into only an elementary school. That building would require some repairs, but most of the work would be things like lowering the water fountains and cabinets, and getting smaller furniture.Holmes said he’s not sure what the grade breakdown of the two schools would be, should the bond pass. That will be determined by the expected growth coming into the district. The district has seen a 19.8 percent increase in enrollment in the last seven years, and with numerous housing developments under or nearing construction in the city, the school district could see an influx on another 700-plus students in the future.
Liverpool manager Jurgen Klopp discussed his midfield options ahead of their Premier League showdown with Manchester United on Sunday.The Reds were the last team to inflict a league defeat on United in what was Jose Mourinho’s final game in charge in December at Anfield in a 3-1 win.Now Liverpool will be looking for a similar result in order to return to the top of the Premier League table in what will be the first of three fixtures in the space of seven days.Therefore, rotations will be expected in the Liverpool team over the course of next week with Klopp able to count on plenty of options in his midfield.“I expect that the players just understand it. It’s a typical thing,” said Klopp on the club website.“I will not explain to everyone why they don’t start because they will all start games.“We don’t have to talk about Ox [Alex Oxlade-Chamberlain] at the moment because he needs a bit of time. He looks so good in training but we need to make him resistant for all the things that are coming up.“For all the others I expect them to understand it. You cannot always have three players in midfield and they play and I have to wrap them in cotton wool – ‘Oh come on please, don’t do this step and this step’ or ‘Oh you feel something, come off training.’“Now we have five or six players for midfield and that’s exactly how it should be.“Of course they cannot start all the time. If somebody makes a story of that, I’m not interested.Liverpool legend Nicol slams Harry Maguire’s Man United form Andrew Smyth – September 14, 2019 Steve Nicol believes Harry Maguire has made some “horrendous mistakes” recently, and has failed to find his best form since joining Manchester United.“If somebody is not happy with that, I’m OK but you have to be ready for the next game and the next game is coming on Wednesday. That’s just how it should be.“All credit to the boys because they always deal with it. I’m not in doubt that they will deal with that situation as well. Make yourself ready because we need to have these different options.”As for Sunday’s match, Virgil van Dijk will make a welcome return to the Reds line-up at Old Trafford after being suspended from Tuesday’s 0-0 Champions League draw against Bayern Munich.Dejan Lovren remains sidelined with a hamstring injury, but Xherdan Shaqiri could feature after making some progress in his recovery from a minor abdominal strain.The match between United and Liverpool will begin at 3:05 PM (CET) on Sunday.I M P A C T 🙌A sublime double from @XS_11official when we hosted @ManUtd earlier this season. 😍 pic.twitter.com/gZzyBwwMeI— Liverpool FC (@LFC) February 22, 2019