<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Natalie Kuster on Medium]]></title>
        <description><![CDATA[Stories by Natalie Kuster on Medium]]></description>
        <link>https://medium.com/@nataliekuster?source=rss-b8fca9187c64------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*tNPAs2Ch61VTt_ki-ZlDpQ@2x.jpeg</url>
            <title>Stories by Natalie Kuster on Medium</title>
            <link>https://medium.com/@nataliekuster?source=rss-b8fca9187c64------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 18:47:08 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@nataliekuster/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Knowledge Representation and Reasoning with Answer Set Programming]]></title>
            <link>https://medium.com/data-science/knowledge-representation-and-reasoning-with-answer-set-programming-376e3113a421?source=rss-b8fca9187c64------2</link>
            <guid isPermaLink="false">https://medium.com/p/376e3113a421</guid>
            <category><![CDATA[editors-pick]]></category>
            <category><![CDATA[answer-set-programming]]></category>
            <category><![CDATA[knowledge-representation]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[digital-transformation]]></category>
            <dc:creator><![CDATA[Natalie Kuster]]></dc:creator>
            <pubDate>Fri, 24 Jul 2020 20:09:56 GMT</pubDate>
            <atom:updated>2020-09-27T15:40:43.900Z</atom:updated>
            <content:encoded><![CDATA[<h4>.. or why penguins should not fly.</h4><p><em>Read about the difference between declarative and imperative programming and learn from code examples (Answer Set Programming, Python and C).</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Vt3T6Fd-5cNynFky" /><figcaption>Penguin ‘Roger’ — slightly bored because he cannot fly. (Photo by <a href="https://unsplash.com/@66north?utm_source=medium&amp;utm_medium=referral">66 north</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><h3>Introduction</h3><p>The amount of computational problems seems to be unlimited in both industry and science. There is a huge demand for new insights from the vast amount of available data. To obtain this knowledge, dedicated people use all kinds of programming languages for designing and implementing algorithms.</p><p>However, many of the current real-world problems are complex (combinatorial) search problems with which we try to automate and optimize processes as well as support people in their decisions. This does not involve encoding of algorithms but encoding <em>given knowledge</em>. In other words, given knowledge about all the rules and constraints to be considered to what is counted as a solution. Instead of writing statements describing the control flow of a computation, declarative programming expresses its logic. <em>What</em> do we want to achieve, not statements about <em>how</em> we can achieve it.</p><p>This is called declarative programming.</p><h3>Why declarative programming?</h3><p>Using a promising declarative programming paradigm, namely Answer Set Programming (ASP, sometimes also Answer Set Prolog), offers unexpected advantages over the popular imperative programming approach, for example:</p><ul><li>Short solutions (as measured by lines of code)</li><li>Transparency (including readability)</li><li>Reliability</li></ul><p>and many more. Once you have broken the problem down into its smallest pieces and written down all the necessary knowledge, you will not only solve the computational challenge. A big advantage is that you have digitized the knowledge and can use it for further problems or make it available in other ways.</p><blockquote>Experience and knowledge can be sustainably secured and protected against loss (e.g. through retirement or termination of employees).</blockquote><p>Furthermore, it strengthens the cooperation between ‘business people’ and programmers. The integration of artificial intelligence is supported and acceptance at all levels will increase. This is a fundamental process that can only be achieved in cooperation with all departments of a company.</p><h3>What is ASP and how does it work?</h3><p>An answer set program consists of given knowledge formulated as facts, rules (head(X) :- body(X).) or constraints. The program is loaded by a solver and returns a “stable model”. This so called <em>answer set</em> consists of all facts that can be derived using the given rules and constraints. A finite amount of stable models are generated as solutions, of which one is finally selected.<br><strong><em>Note: </em></strong><em>grounding and solving the problems are not within the scope of this article.</em></p><p>Let us consider a famous example from logic about birds and penguins. It is a well known fact that birds can fly. This can be encoded as an answert set rule:</p><pre>canFly(X) :- bird(X).</pre><p>The rule is read as <em>“If </em><em>X is a bird, then </em><em>X can fly”</em>. Now let’s add more knowledge! For example, facts that tell the unconditional truth, just as seagull ‘Malvin’ is a bird, but also penguin ‘Roger’ is one.</p><pre>canFly(X) :- bird(X).<br>bird(malvin).<br>bird(roger).</pre><pre>== MODEL OUTPUT ==<br>Anwer: 1<br>bird(malvin) bird(roger) canFly(malvin) canFly(roger)<br>SATISFIABLE</pre><p>As a result, the answer set tells us the known facts that Malvin and Roger are birds but also concluded that they are able to fly. The biology enthusiasts among us know, that penguins are unable to fly thus the model is wrong!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*-a2Z4y7ZnQ4Y83Cw" /><figcaption>Seagull ‘Malvin’ — slightly showing off because he can fly. (Photo by <a href="https://unsplash.com/@philbotha?utm_source=medium&amp;utm_medium=referral">Phil Botha</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><p>In order to get acceptable results we need to add more knowledge in form of facts and integrity constraints to the program. Here, the model needs to know that Roger is not only a bird but also a penguin and that there is no bird that is a penguin which is able to fly.</p><pre>canFly(X) :- bird(X).<br>bird(malvin).<br>bird(roger).<br>seagull(malvin).<br>penguin(roger).<br>:- canFly(X), penguin(X).</pre><pre>== MODEL OUTPUT ==<br>UNSATISFIABLE</pre><p>This does not bring the expected result and shows how important it is to think thoroughly and very carefully about the problem as well as the solution. The model tells, like stated by the programmer, that a bird can fly but there is no advise for birds that belong to the penguin family. To avoid this, we could add that only birds that are not penguins are able to fly.</p><pre>canFly(X) :- bird(X), not penguin(X).<br>bird(malvin).<br>bird(roger).<br>seagull(malvin).<br>penguin(roger).<br>:- canFly(X), penguin(X).</pre><pre>== MODEL OUTPUT ==<br>Answer: 1<br>bird(malvin) bird(roger) seagull(malvin) penguin(roger) canFly(malvin)<br>SATISFIABLE</pre><p>Adding this knwoledge in Line [1], the stable model consists of the expected outcome.</p><h3>Imperative versus declarative programming</h3><h4>.. for solving Sudoku</h4><p>Let’s look at another example to highlight the advantages of ASP mentioned above. Sudoku is a well known puzzle game and popular for explaining search problems. Given an initial 9x9 grid of cells containig numbers between 1 and 9 or blanks, all blanks must be filled with numbers. You win Sudoko if you find all values such that every row, column, and 3x3 subsquare contains the numbers 1–9, each with a single occurrence.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LBBgLEQLc-NEIjrK6Hd3xQ.png" /><figcaption>Exemplary Sudoku game (adopted from Potassco.org).</figcaption></figure><h4>Python (imperative)</h4><p>The following Code snippet will show how to solve the Sudoku in Python.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Frepl.it%2F%40nkuster%2Fmediumsudokupython%3Flite%3Dtrue&amp;display_name=repl.it&amp;url=https%3A%2F%2Frepl.it%2F%40nkuster%2Fmediumsudokupython%3Flite%3Dtrue&amp;image=https%3A%2F%2Frepl.it%2Fpublic%2Fimages%2Freplit-logo-800x600.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=repl" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/dcd6d2f09f7b441584cc33b87d9d111f/href">https://medium.com/media/dcd6d2f09f7b441584cc33b87d9d111f/href</a></iframe><h4>C (imperative)</h4><p>Solving the Sudoku game in C looks very similar to Python. There is no significant difference in the approach.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Frepl.it%2F%40nkuster%2Fmediumsudokuc%3Flite%3Dtrue&amp;display_name=repl.it&amp;url=https%3A%2F%2Frepl.it%2F%40nkuster%2Fmediumsudokuc%3Flite%3Dtrue&amp;image=https%3A%2F%2Frepl.it%2Fpublic%2Fimages%2Freplit-logo-800x600.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=repl" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/a78029090464292202b6907676dc490e/href">https://medium.com/media/a78029090464292202b6907676dc490e/href</a></iframe><p>Both, Python and C, show written statements describing the control flow of a computation, thus, the ‘<em>how to solve it’ </em>and how the state has to be changed for the upcoming step.</p><h4>ASP (declarative)</h4><p>Last but no least, let’s have a look at the ASP code snippet — after initalization, we are able to solve the Sudoku with<strong> less than 10 lines of code</strong>!</p><pre>% Initialize the game (given numbers)<br>sudoku(1, 1, 5).<br>sudoku(1, 2, 3).<br>sudoku(1, 5, 7).<br>sudoku(2, 1, 6).<br>sudoku(2, 4, 1).<br>sudoku(2, 5, 9).<br>sudoku(2, 6, 5).<br>sudoku(3, 2, 9).<br>sudoku(3, 3, 8).<br>sudoku(3, 8, 6).<br>sudoku(4, 1, 8).<br>sudoku(4, 5, 6).<br>sudoku(4, 9, 3).<br>sudoku(5, 1, 4).<br>sudoku(5, 4, 8).<br>sudoku(5, 6, 3).<br>sudoku(5, 9, 1).<br>sudoku(6, 1, 7).<br>sudoku(6, 5, 2).<br>sudoku(6, 9, 6).<br>sudoku(7, 2, 6).<br>sudoku(7, 7, 2).<br>sudoku(7, 8, 8).<br>sudoku(8, 4, 4).<br>sudoku(8, 5, 1).<br>sudoku(8, 6, 9).<br>sudoku(8, 9, 5).<br>sudoku(9, 5, 8).<br>sudoku(9, 8, 7).<br>sudoku(9, 9, 9).<br></pre><pre>% define the grid<br>n(1..9).<br>x(1..9).<br>y(1..9).</pre><pre>% each field contains exactly one number from 1 to 9<br>{sudoku(X,Y,N): n(N)} = 1 :- x(X) ,y(Y).</pre><pre>% helper<br>subgrid(X,Y,A,B) :- x(X), x(A), y(Y), y(B),(X-1)/3 == (A-1)/3, (Y-1)/3 == (B-1)/3.</pre><pre>% constraints<br>:- sudoku(X,Y,N), sudoku(A,Y,N), X!=A.<br>:- sudoku(X,Y,N), sudoku(X,B,N), Y!=B.<br>:- sudoku(X,Y,V), sudoku(A,B,V), subgrid(X,Y,A,B), X != A, Y != B.</pre><pre>#show sudoku/3.</pre><pre>== MODEL OUTPUT ==<br>Answer: 1<br>(&#39;sudoku&#39;, (2, 1, 6)), (&#39;sudoku&#39;, (1, 2, 3)), (&#39;sudoku&#39;, (9, 4, 2)), (&#39;sudoku&#39;, (7, 2, 6)), (&#39;sudoku&#39;, (7, 3, 1)), (&#39;sudoku&#39;, (1, 9, 2)), (&#39;sudoku&#39;, (5, 1, 4)), (&#39;sudoku&#39;, (7, 4, 5)), (&#39;sudoku&#39;, (4, 3, 9)), (&#39;sudoku&#39;, (9, 1, 3)), (&#39;sudoku&#39;, (4, 5, 6)), (&#39;sudoku&#39;, (5, 6, 3)), (&#39;sudoku&#39;, (2, 4, 1)), (&#39;sudoku&#39;, (8, 1, 2)), (&#39;sudoku&#39;, (8, 8, 3)), (&#39;sudoku&#39;, (7, 9, 4)), (&#39;sudoku&#39;, (8, 7, 6)), (&#39;sudoku&#39;, (5, 4, 8)), (&#39;sudoku&#39;, (7, 6, 7)), (&#39;sudoku&#39;, (8, 6, 9)), (&#39;sudoku&#39;, (6, 5, 2)), (&#39;sudoku&#39;, (9, 7, 1)), (&#39;sudoku&#39;, (3, 4, 3)), (&#39;sudoku&#39;, (4, 7, 4)), (&#39;sudoku&#39;, (3, 3, 8)), (&#39;sudoku&#39;, (4, 8, 2)), (&#39;sudoku&#39;, (1, 7, 9)), (&#39;sudoku&#39;, (9, 9, 9)), (&#39;sudoku&#39;, (9, 8, 7)), (&#39;sudoku&#39;, (5, 9, 1)), (&#39;sudoku&#39;, (4, 4, 7)), (&#39;sudoku&#39;, (6, 9, 6)), (&#39;sudoku&#39;, (7, 7, 2)), (&#39;sudoku&#39;, (2, 7, 3)), (&#39;sudoku&#39;, (5, 5, 5)), (&#39;sudoku&#39;, (1, 5, 7)), (&#39;sudoku&#39;, (1, 1, 5)), (&#39;sudoku&#39;, (6, 3, 3)), (&#39;sudoku&#39;, (2, 6, 5)), (&#39;sudoku&#39;, (2, 9, 8)), (&#39;sudoku&#39;, (8, 3, 7)), (&#39;sudoku&#39;, (3, 6, 2)), (&#39;sudoku&#39;, (3, 8, 6)), (&#39;sudoku&#39;, (2, 8, 4)), (&#39;sudoku&#39;, (1, 3, 4)), (&#39;sudoku&#39;, (8, 2, 8)), (&#39;sudoku&#39;, (3, 7, 5)), (&#39;sudoku&#39;, (9, 5, 8)), (&#39;sudoku&#39;, (4, 9, 3)), (&#39;sudoku&#39;, (6, 4, 9)), (&#39;sudoku&#39;, (7, 5, 3)), (&#39;sudoku&#39;, (2, 5, 9)), (&#39;sudoku&#39;, (8, 5, 1)), (&#39;sudoku&#39;, (5, 3, 6)), (&#39;sudoku&#39;, (4, 6, 1)), (&#39;sudoku&#39;, (3, 9, 7)), (&#39;sudoku&#39;, (1, 4, 6)), (&#39;sudoku&#39;, (9, 2, 4)), (&#39;sudoku&#39;, (5, 8, 9)), (&#39;sudoku&#39;, (1, 8, 1)), (&#39;sudoku&#39;, (6, 6, 4)), (&#39;sudoku&#39;, (5, 7, 7)), (&#39;sudoku&#39;, (7, 1, 9)), (&#39;sudoku&#39;, (3, 5, 4)), (&#39;sudoku&#39;, (6, 8, 5)), (&#39;sudoku&#39;, (5, 2, 2)), (&#39;sudoku&#39;, (2, 3, 2)), (&#39;sudoku&#39;, (8, 9, 5)), (&#39;sudoku&#39;, (9, 6, 6)), (&#39;sudoku&#39;, (9, 3, 5)), (&#39;sudoku&#39;, (6, 2, 1)), (&#39;sudoku&#39;, (3, 1, 1)), (&#39;sudoku&#39;, (4, 2, 5)), (&#39;sudoku&#39;, (6, 1, 7)), (&#39;sudoku&#39;, (4, 1, 8)), (&#39;sudoku&#39;, (8, 4, 4)), (&#39;sudoku&#39;, (2, 2, 7)), (&#39;sudoku&#39;, (3, 2, 9)), (&#39;sudoku&#39;, (1, 6, 8)), (&#39;sudoku&#39;, (6, 7, 8)), (&#39;sudoku&#39;, (7, 8, 8))}<br>SATISFIABLE</pre><p><strong><em>Note</em></strong><em>: using for example a Python wrapper for ASP, the result can look as handy as in the code examples above.</em></p><p>Compared to Python and C, this is all about the ‘what’ and not about the ‘how’ — a fundamental difference between the two approaches.</p><p>At the very beginning, we define that the Sudoku board is 9x9 cells (or fields) and we use the numbers from 1 to 9 as the values to fill in. Following the basic rule that only one number between 1 and 9 may be filled in each field, we define a little helper. It states which fields refer to the same <em>sub-grid</em>. Finally, we only need the constraints to tell the solver what is possible and what is not. The first line checks, whether the choosen number is unique in the column whereas the second constraint checks the rule for rows. The third constraint completes the rules of the game, according to which a number must also be unique in the subgrid.</p><p>That was it. I admit the syntax takes getting used to. But once you’re familiar with it, you can create incredibly readable programs that solve highly complex problems for us.</p><h3>Summary</h3><p>ASP is a very promising tool for knowledge preservation and declarative problem solving in the area of Knowledge Representation and Reasoning.</p><p><strong>Short solutions</strong> — Apart from the initial effort to map the Sudoku game, ASP provides by far the shortest way (measured in lines of code) to the solution.<br><strong>Transparency</strong> — There is no need to create rules which are unclear to the user and which concern the status of the programme. Only the basic conditions and the three rules of the game must be coded in order to end up at the correct result.<br><strong>Reliability</strong> — All rules of the game are fixed and the solver cannot create a black box with solution steps that are not traceable afterwards. This may not be important in the example given, but it is a major issue in many industrial applications.</p><p>To mention only a few real-world examples, decision support for space shuttles at NASA, train scheduling in Switzerland — one of the world’s densest train network— , or team building at the largest transshipment terminal on the Mediterranean coast are evidence of its potential.</p><h4>Bottom Line</h4><p>The acceptance of artificial intelligence in industry is still very low. In addition, many companies are simply overwhelmed by the existing flood of data.</p><blockquote>Bring the knowledge of operational subject matter experts together with the strengths of Artificial Intelligence.</blockquote><p>ASP offers an unique opportunity to digitalize and protect existing knowledge. Due to the good readability and the required understanding of the process as well as the problem to be solved, IT and business move closer together and thus achieve better and more sustainable success.</p><h4>Did you like the story?</h4><p>I appreciate likes and retweets of this article – there will be more about this topic soon! You can also find an executive summary <a href="https://towardsdatascience.com/how-machine-learning-made-ai-forget-about-knowledge-representation-and-reasoning-cbec128aff56">here</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=376e3113a421" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/knowledge-representation-and-reasoning-with-answer-set-programming-376e3113a421">Knowledge Representation and Reasoning with Answer Set Programming</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Machine Learning made AI forget about Knowledge Representation and Reasoning]]></title>
            <link>https://medium.com/data-science/how-machine-learning-made-ai-forget-about-knowledge-representation-and-reasoning-cbec128aff56?source=rss-b8fca9187c64------2</link>
            <guid isPermaLink="false">https://medium.com/p/cbec128aff56</guid>
            <category><![CDATA[knowledge-representation]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[editors-pick]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[logical-reasoning]]></category>
            <dc:creator><![CDATA[Natalie Kuster]]></dc:creator>
            <pubDate>Mon, 02 Mar 2020 21:00:16 GMT</pubDate>
            <atom:updated>2020-08-14T07:38:01.283Z</atom:updated>
            <content:encoded><![CDATA[<h4>A brief history of how I learned about the forgotten core of artificial intelligence.</h4><p>In my early days working as a data scientist in AI, I was taught one thing above all: you need more and better data to feed your learning systems. And I have dedicated myself to finding that data.</p><p>I was lured into the world of machine learning while trying to discover the world of artificial intelligence. I admit, it is exhilarating to make a computer do complex things to my liking without saying <em>how</em> it should do this. I ran riot in this world and have long forgotten my original goal whilst studying how to prepare data, engineer features and build deep networks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*DxdcJnnORwSluMrWJ_LpDQ.jpeg" /><figcaption><a href="https://media.makeameme.org/created/machine-learning-machine-884d19c7bd.jpg">https://media.makeameme.org/created/machine-learning-machine-884d19c7bd.jpg</a></figcaption></figure><p>However, in some cases where I was asked to automate and optimize a business problem, the issue of the availability, quantity and quality of associated data often remained wanting. In addition, the increased acceptance of artificial intelligence in the industry gave rise to fear, skepticism and resentment. People think they are being replaced by a black box. That is not the truth. But I do understand the resistance when you try to replace years of expert knowledge with an opaque box.</p><blockquote>Artificial intelligence is commonly described as the discipline of computer perception, reasoning, and action.</blockquote><blockquote>I missed the reasoning.</blockquote><p>John McCarthy coined the term ‘artificial intelligence’ in 1955 and organized the <a href="https://en.wikipedia.org/wiki/Dartmouth_workshop">Dartmouth Research Project </a>in Summer 1956 — the start of AI as a field.<br>He stated:<em> “The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.”</em></p><p>He developed <a href="https://en.wikipedia.org/wiki/Lisp_(programming_language)">Lisp</a> in 1958 and reinvented how we program to create thinking machines. And others followed him, e.g. Alain Colmerauer and Philippe Roussel, who came up with <a href="https://en.wikipedia.org/wiki/Prolog">Prolog</a> in 1972. Prolog was partly motivated by the desire to reconcile the use of logic as a declarative knowledge representation language with the procedural representation of knowledge.<br>Thus, Knowledge representation and reasoning (KRR) became a key area of AI concerned with how knowledge can be represented symbolically and manipulated in an automated way by reasoning programs. In other words, how an agent uses what it knows in deciding what to do.</p><p>Consider the task of baking a cake. Of course, I could collect data on the amount of ingredients and the order in which they are usually mixed to achieve the best results. But I like Grandmother’s cake best anyway and I know the recipe very well. So why shouldn’t I explain to the system right away which rules have to be observed when mixing the ingredients and which basic conditions have to be created? The same happens when we study complex processes in industry. Too often, people first reflexively ask for existing data and weigh up what additional data should be collected for a successful project. Only a few shrewd people are interested in the details of the processes underlying the problem. It often turns out that the automatization or optimization is just a combinatorial problem, albeit a very complex one.</p><blockquote>When I think of a new problem, my first question no longer refers to the data, but to the problem itself.</blockquote><p>The development of a process, often over many years and enriched with a lot of experience and expert knowledge, makes it increasingly complex. New dependencies are being added and the constantly growing requirements demand more precise and above all faster results. A task that can simply no longer be accomplished by the human brain. With the digital and systematic recording of the existing knowledge as well as the detailed process steps and constraints, KRR can not only automate the process but also optimize it. Examples are work shift planning, radio frequency auction, time tabling and many more.</p><blockquote>Learning from data is extremely popular — but not your only option.</blockquote><p>With all the knowledge of machine learning and looking back at what has already been achieved in the field of AI, we should always use one of our strongest and most human abilities: <strong>creativity</strong>. Think outside the box and do not choose a method because it sounds best or attracts attention. Choose the method that best fits your problem.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*kM3SnYe4XRqH6gSi" /><figcaption>Photo by <a href="https://unsplash.com/@askkell?utm_source=medium&amp;utm_medium=referral">Andy Kelly</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h4><strong>Bottom line</strong></h4><p>Do not let a black box naturally decide how the processes are designed. Talk to the experts and map their process. With the necessary transparency we create trust and pave the way for a higher acceptance of artificial intelligence within society.</p><p>Learn more about Knowledge Representation and Reasoning with Answer Set Programming in <a href="https://towardsdatascience.com/knowledge-representation-and-reasoning-with-answer-set-programming-376e3113a421">this article</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cbec128aff56" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/how-machine-learning-made-ai-forget-about-knowledge-representation-and-reasoning-cbec128aff56">How Machine Learning made AI forget about Knowledge Representation and Reasoning</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>