Yann LeCun – Power & Limits of Deep Learning

I want to share with you one of the most accessible and informative videos I’ve seen so far on deep learning. What really surprised me is the lower number of views compare to others. Future based on facts might not be as appealing as Science Fiction or alarming videos.

Yann LeCun is Director of AI Research at Facebook, and Silver Professor of Dara Science, Computer Science, Neural Science, and Electrical Engineering at New York University, affiliated with the NYU Center for Data Science, the Courant Institute of Mathematical Science, the Center for Neural Science, and the Electrical and Computer Engineering Department.

Video From Prof. Harari

I will begin my series of articles on Artificial Intelligence by commenting a video from Prof. Yuval Harari. He´s a historian, philosopher and not from the technical world which make it very interesting because he is seeing AI from an external point of view.

The main ideas that attacked my attention in this video are:

  • Ai is all about statistical calculation and probability,
  • Actual result obtain with AI are acceptable and will get better in the future because more and more data will be recording.
  • People will soon think that computer knows better and will start believing that it is always correct.
  • Include human biology into human behaviour prediction is the next future revolution.
  • Conclusion: Computer will soon be able to know what you want, better than you,

The presenter is right in explaining that we are talking about statistical calculation and probability. We are calculating probabilities based on the user’s data history and / or comparison with other people who exhibit the same behavioral characteristics than him. The result is good but far to be perfect as Prof Harari recognized.

I have an issue with his reasoning. He believes AI technology will improve and the errors in predictions that we have nowadays will finally become negligible.  Computer will soon be able to know what you want, better than you without considering that these errors are intrinsic of this technology: using the past to forecast the future and a human is a complex system.

Do you remember the butterfly effect? A very small change in initial conditions had created a significantly different outcome. This is due to the difficulties to describe well enough the system at one time and its sensitive dependence on initial conditions. We are collecting huge amount of weather data and process them through supercomputers but I still don´t know if I can organise a barbecue in 4 days and be sure that it will not rain.

The danger is not that the computer will know better than us. Prof Harari mentioned in his video and I fully agree with him.  The danger is that we start to believe that the computer knows better than us and stop questioning the result of its calculation.


Why I want to write about Artificial Intelligence (AI)

I want to explain why I will post several articles about Artificial intelligence.

Before my friend Marwan ask me to join him in Cadmus, I had basic computer skills. I´m engineer but computer science wasn´t my field. Also, I was so busy, I was listening to the marketing talks around Artificial Intelligence, but never really thought about it. It was a black box.

Everything changed with the success of our plug-in “Today for Outlook”. We decided to create new functions to help people even more: The solution to reduce the number of repetitive or boring tasks or just simply make life easier is to use Artificial Intelligence technology. So, I started to learn about Artificial Intelligence.

My goal was not to become a data scientist. I wanted to understand what is possible and most importantly, the problems and limitations of this technology. That’s what I want to share with you in the coming articles.

Start the day with ZERO emails, thanks to Cinderella’s magic – Robert´s feedback.

Robert is used to keep an eye on his mailbox all the time, even while in meetings. What problem does he have? On several occasions in the past, he forgot to answer important emails after his meetings.

“I know it’s a bad habit, but I have to watch my mail all the time, even when I’m in a meeting. I learned the hard way, I cannot count on my memory. I do not count the times I remembered too late that I had to answer an important email. Or even worse, when I completely forgot to answer one.  I have tried using different techniques like changing the email to unread, or creating a task, or taking a note, etc. But there was a lack of consistency and it still happened. Someone distracts me, asks me a question at the wrong time and I don’t use the technique. You know Murphy´s law! It always happens to the one email you must not miss. I needed a better solution and for me, “Today for Outlook” is doing the job.

Now, I don´t think anymore which technique is the most appropriate. When I see an important email, I send it to Today folder in one click. After the meeting, I check my TODAY folder. It has become a routine. No more late answers, no more forgotten email and I didn´t have to change my bad habit, I can still look at my emails during meetings. It´s make me happy”

Continue sharing with us how you use this plug-in. The range of situations where this little plug-in helps is just amazing and doesn´t stop to surprise us.

General presentation

Instead of working in your overloaded inbox, start from ZERO e-mail in the TODAY folder. Why this different approach?

You will take back control over your emails by deciding which ones deserve your attention and become special ones. Remember Cinderella´s magic will stop working at midnight. All these special emails contained in TODAY folder will come back to normal emails and will be sent back to your Inbox.

TODAY folder is now empty and ready for another magical day.


Use the following link to go to Microsoft Official Website (this add-in has been tested and approved by Microsoft).


Start the day with ZERO emails, thanks to Cinderella’s magic – Marc´s Feedbacks

Marc is used to quickly browse through his e-mails in order to answer them in order of urgency and importance. What is his problem?

So many time when I work on my mail, I get interrupted or called to solve an emergency and I forget to answer some important emails. I found a solution but far from optimum: I  open the most important emails and left them opened, but quickly it is becoming a mess with dozens of windows opened.”

Off course, we asked him if our free plug-in “Today for Outlook” helps him.

I haven´t changed my habits, I browse quickly my emails but now I send all the important ones to the TODAY directory. After getting disturbed, I cannot avoid it ☹, I know immediately how many important e-mails are left but most important, I don´t forget any of them anymore. And this affected positively my stress level and my behavior with colleagues.”

Continue sharing with us how you use this plug-in. The range of situations where this little plug-in helps is just amazing and doesn´t stop to surprise us.

General presentation

Instead of working in your overloaded inbox, start from ZERO e-mail in the TODAY folder. Why this different approach?

You will take back control over your emails by deciding which ones deserve your attention and become special ones. Remember Cinderella´s magic will stop working at midnight. All these special emails contained in TODAY folder will come back to normal emails and will be sent back to your Inbox.

TODAY folder is now empty and ready for another magical day.


Use the following link to go to Microsoft Official Website (this add-in has been tested and approved by Microsoft).


Process review of Today app.

Actions of the plug-in (without human method and process):

  • When press of the icon TODAY, the following is happening:
    • The email will be moved to “TODAY” directory (created at the installation of the plug-in)
    • The email is marked as unread.
  • During the night,  the following is happening:
    • Any email contained in Today directory is moved back to inbox
    • The email is marked as unread.
    • TODAY directory is now ready for next day.

From a process point of view:

Now it is important to think how this relatively simple tool is affecting the user behaviour and routine and what would be the output.

We have created a character called Archie so we could use him to look at the situation before and after.

Archie needs to find the time to treat his emails which never stop to come. Other events and activities are also competing for Archie´s time. To make it worse, some of these events or activities were not plan or known so his time schedule could change anytime. Finally, Archie is now accessing his emails from different devices, in different location with more or less time available.

Before using this plug-in:

Archi had several choices when looking at an email. He could:

  • Answer it immediately,
  • Just read it because it is for information only,
  • Delete it,
  • Leave it for later. The ISSUE is that Archi has to remember it. He might be so scared to forget it that Archie decide to answer it immediatly.

Many ways to remember an email already exist:

  • Archi can change back the status of the email to unread to open it again. The problem with this option is that this email is just one within all the other unread email.
  • Archie can flag it to highlight this email but he would need some time to enter all the relevant information so realistically, Archie is not doing it often.
  • Another option is that Archie can send this email back to himself as important immediately or at a later time but he is not sure he will select the right time.
  • Archie can also create a task, which is not very fast
  • Archie can just take a note on your agenda (the good old way)
  • etc…

If you think about it, every time that Archie read an important emails but not urgent and would like to leave it for later, he needs to think and decide which option he will select. The number of options makes that Archie is not very consistent in his way he treats “important but not urgent emails”. Archie does not realise that he is making his work harder because now he has “important but not urgent emails” in different folders, with different flag, some as tasks, some as note on his agenda…

Using this plug-in:

Now, in one click, you are:

  • Highlighting this emails as important but not urgent,
  • Moving it to the TODAY folder so they are separated from the other emails.
  • You will better organise your time as can see in one glance how many unread email are in the Today folder.

Why we bring emails contained into TODAY folder back to inbox?

  • Important emails need to be archived and this is what we expect to happen. The number of emails contained into TODAY folder should decrease during the day to reach zero when Archie leaves office.
  • Any process need some type of constrain to stop complacency and Archie deviating from it. In our case, it is a soft constrain as we believe that nobody likes redoing the same operation twice as Archie will need to resend the email to TODAY folder. But not only!!!

Why emails contained into TODAY folder are moved back to inbox as unread emails?

  • It is also a security for Archie. Maybe, he was busy this day and he didn´t treat all the important emails contained into TODAY folder so they might become even urgent. Human nature makes that we start always by looking at the inbox to see what´s new and Archie is not different. Only later, he would think to go to our TODAY folder and it might be too late! So better to find the email where Archie would look first.
  • Maybe Archie didn´t finish working on an email. As he opened it, its status would be “read” but if we move it back to his inbox without changing its status to unread, Archie might forget about it or might lose time looking for it. By changing it to unread, Archie will not miss any important emails.

This simplicity and the repeatability in the way Archie would be treating emails which are important but not urgent makes the difference. Now Archie has a solid process.

Not only Archie has a solid process, he is also saving time by avoiding multiple clicks or taking unnecessary decisions. All this saving add up very quickly.  

What is Process Mining?

Process mining is a technique that facilitates the analysis of business processes by using event logs that can be extracted from computer systems. The objective is the acquisition of knowledge of real processes from the processing of a large amount of data, by automatic or semi-automatic methods.

The strong point of process mining is to be able to observe the behavior of the organization, based on facts and actual figures over a period of time. We therefore have a solid foundation for conducting a thorough study of business processes and their improvement. Nevertheless, the result obtained using process mining software would need to be combined with human investigation to fully understand the situation.

Main applications of process mining

Process Mining essentially supports three process management disciplines:

Business Process discovery: which consists of the graphic representation – thus the reconstruction based on event logs – of the current processes of the organization; including possible process variants. The result is a process model – or even an animated diagram.

Conformance checking: which is the comparison of an existing process model to an event log of the same process, in order to verify if the model corresponds to reality (as recorded in the log), or vice versa, if the Implementation of the process is well in line with the model. This makes it possible to check if the (company) rules have been respected or not.

Business Process enhancement & improvement: Thanks to the graphs and measurements obtained from Process Mining algorithms, it is easier to detect bottlenecks, to see which resources are underutilized, and so on. The many possible points of view of such an analysis facilitate the detection of the causes of inefficiency or errors within a process and thus the organization.

Advantages of process mining

Here are the main advantages of Process Mining, according to the study of J. Claes and G. Poels:

Objectivity: facts do not lie. The Event Logs (group of actions and date recorded by the system) reflect how a process is actually working. Models reconstructed using Process Mining are not influenced by a human perception that is often subjective.

This is the reason why you will have differences between models obtained from process mining and the ones obtained through interviews or workshops. But there are also information to learn here!!!

Speed: Even if you still need to do your homework and to have an understanding of the context in advance (to select which information are relevant and which Event Logs to use), it is obvious that process representation through Process Mining speeds up mapping and modeling activities.

Less effort (more efficiency): Process discovery through process mining is efficient. You will minimize the number of interviews or workshops with business people and they will be more efficient as you will review and discuss facts.

Full state: One very important point, the analysis of relevant event logs makes it easier to distinguish between the main process and other variants or exceptions if you run the analysis over a sufficient period of time. You can spot variants and exceptions that you might not discover without Process Mining because people forgot to mention them in the interviews

Transparency: we are looking at facts recorded by the system and if needed, we can zoom to details as who performed a transaction, at what time, and so on. This can help to get more transparency on how the organization really works.

Compliance: As explained in “full state”, process mining helps identifying variant, exceptions from the main process. Also this main process can be compared to the desired or defined process to obtain nonconformities.

Root-causes & bottlenecks: visualization of how the process is running helps discovering root causes, bottlenecks, etc . Our objective is now to look how to improve the processes.

Predictions & simulations: the facts helped us to build a model. Now we can modify the model and see how it will perform running the data. This simulation makes it possible to predict the future behavior of a process with the modification we are proposing.

Challenges of process mining

Despite the many benefits of Process Mining, it also presents challenges and you need to be aware of them. The challenges are very similar with the challenges encountered in data mining and business intelligences. Claes & Poels mention the following challenges:

Data quality: Data quality is a full subject in itself and it is not the subject of this post. Building the Events log might require the use of different sources which have different formats, contain errors, etc.  This bring us to the next point.

Cost of data preparation: it might take effort to clean the data and this represent a cost.

Access to data: It is not always possible to obtain valid data. Especially when dealing with ‘older’ computer systems that do not produce Event Logs.

Technology: algorithms have limitation and are not suitable to deal with all possibilities and situations which could occur during the process execution.

Usability of the tool: there are different providers and process mining software present different combination of flexibility and user-friendliness. More flexibility comes with more set up to be performed but you can performed more type of analysis.

Cultural Intelligence

Both, Marwan and Thierry, have an extensive international experience. We were living and working in many different countries and we know that is very important to take into account cultural differences when managing people and in implementing or adapting processes. What works in one countries might present some challenges somewhere else.

We are presenting a video so the people which are not familiarized with cultural differences will understand better the problem.

The situation comes from our own experience. Cultural differences make the story funny due to incomprehension, miss understanding between the 2 sides, but remember, there is something to learn . . .  From a company perspective, cultural differences could have  catastrophic results.

Have Fun:

Management of Change

The following information comes from:

Sreekanth Ramakrishman, advisory engineer/scientist & learning specialist, IBM Center for Learning and Development.

Michael Testani, learning consultant, portfolio manager & master black belt, IBM Center for Learning and Development.

So according to Ramakrishman and Testani, “Far too often business transformation efforts concentrate on the process improvement strategies and business process reengineering; while essentially ignoring the people aspect of the change initiative. Subsequently, these transformation initiatives do not achieve their desired results. Studies have shown that approximately three quarters of business re-engineering efforts do not achieve their objectives and subsequently do not sustain themselves over the long term, and one of the most commonly cited reasons for their failing is due to the lack of focus on the organization’s culture.” By aligning people, process, and technology, companies can develop “critical organizational competencies around organizational culture transformation and process improvement; resulting in a more effective and sustainable change effort” [Ramakrishman and Testani, 2011].

Say differently:

It would be a big mistake to think that the process you are modifiying or the new process you are implementing, because they will make the work of your employees easier, will be accepted and you will have employees support.

How to manage people through complex changes?

As all manager knows it is not easy to manage people through changes. You will always facing a lot of resistance and you need to take it into account from the start.