<?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[NWiMLDS - Medium]]></title>
        <description><![CDATA[A community for women working and with interest in data science and machine learning - Medium]]></description>
        <link>https://medium.com/nairobiwimlds?source=rss----a090e1dd8a47---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>NWiMLDS - Medium</title>
            <link>https://medium.com/nairobiwimlds?source=rss----a090e1dd8a47---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 17:55:44 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/nairobiwimlds" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[The Explore Series: Meet the Faculty]]></title>
            <link>https://medium.com/nairobiwimlds/the-explore-series-meet-the-faculty-daa63af1a98f?source=rss----a090e1dd8a47---4</link>
            <guid isPermaLink="false">https://medium.com/p/daa63af1a98f</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[africa]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[AI & Insights]]></dc:creator>
            <pubDate>Wed, 15 Apr 2020 07:23:47 GMT</pubDate>
            <atom:updated>2020-04-15T07:23:47.111Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Join us for an exciting 4 month series to explore Data Science, Machine Learning and Artificial Intelligence.</em></p><h4>About the Series:</h4><p>The coursework and sessions are underwritten by industry experts and offer an excellent resources on how to turn ideas, data and models into products that scale. The session will be available live on Saturday, 18th April 2020 through the listed links.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*lgYtECwCvfLIUYT2R7XF-g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*BP64EHVblkQZdggPiinmgQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*0nEUU-Ke5Ag78iDa9S4-8Q.png" /><figcaption>The Explore Series — 2020</figcaption></figure><p>The opening Keynotes facilitated by Everlyn Asiko and Mercy Njue will run from 8:30am to 9:30am and will be available through <a href="https://meet.jit.si/TheExploreSeriesKeynotes">this link.</a></p><p><strong>Everlyn Asiko</strong> is a data scientist with a passion for teaching and mentoring. Her work in accident location extraction from natural language processing aims at drawing location insights from natural text and has been presented at a Pan African AI conference as a poster. She is a Data Science technical mentor at Moringa school and also mentors Google Africa scholars in the Andela Learning Community program. Previously, she worked at Magharibi Innovation Hub as a community manager where she was in charge of the projects and the individuals at the hub.</p><p><strong>Mercy Njue</strong> is the founder and CEO of Botlab, a company that provides training in data science and cyber security. She has a bachelor’s degree in Business and Technology as well as an MBA from Africa Nazarene University. She is trained in General Data Protection Regulation (GDPR) by University of Groningen in the Netherlands and is a certified IBM cloud professional, as well as a graduate of the Data Science for Business Leader program by Data Incubator Chicago. She leads the Nairobi chapter of Women in Data, an organization that supports women exploring careers in data science. Mercy founded Botlab with the mission to highlight the power of technology. Her goal is to increase the uptake of technology use in Kenya. To date, Botlab has trained over 3000 students and professionals. Programs run include courses for company leaders to emphasize the power of data, sessions on cyber security and risk mitigation, and tech master classes for SMEs. Mercy is a strong believer in giving back and opening doors for those that follow in her footsteps. She is a leading voice on the Women in Tech Kenya platform, and has contributed immensely to the advancement of women in STEM.</p><h4>Explore Artificial Intelligence:</h4><p>Facilitated by <a href="https://ke.linkedin.com/in/katie-wanjik%C3%BC">Katie Wanjiku</a>, this session will provide hands-on insights into building self-driving cars with a demo in PX4 Rover kit.</p><p>Katie Wanjiku has built industry-scale AI and robotics for companies in Silicon Valley, London and Bangalore. She is presently engaged at Facebook, Antler and Africa’s Talking Labs focused on hitec gaps in Sub Sahara. Her mission is to power local talents and businesses to scale globally and transfigure the ecosystem for good.</p><p>The session will start at 9:30am and will be available through <a href="https://meet.jit.si/TheExploreSeriesAITrack">this link.</a></p><h4>Explore Data Science:</h4><p>This session, facilitated by <a href="https://twitter.com/MathengeMaureen">Maureen Mathenge</a> and <a href="https://ke.linkedin.com/in/cynthia-asingwa-khazenzi">Cythia Asingwa</a> will cover key foundational concepts into practical data science.</p><p>Maureen is a very enthusiastic individual with a Bachelor of Economics and Statistics.She engaged in a student-run organization,AIESEC,where she nurtured her leadership skills coordinating volunteer and internship programs. She has experience in managing high performing teams,communication and operational planning and implementation. She is enrolled in a data science course where she continuously grows her skill in data analysis and machine learning.</p><p>Cynthia Asingwa is a Data Science junkie with a degree in Mathematics and Computer, majoring in Applied Mathematics. She derives joy in supporting research in maternal and paediatric health care with a goal in learning what solutions make health systems tick in Sub-Saharan Africa. Apart from learning on her own, she enjoys being amongst other learners and believes that knowledge sharing is a powerful tool that allows for the coming up of innovative, inclusive and appropriate solutions to our contextual challenges.</p><h4>Explore Machine Learning:</h4><p>Led by <a href="https://ke.linkedin.com/in/zakia-mustafa">Zakia Mustafa</a> &amp; <a href="https://ke.linkedin.com/in/lydia-gathoni-25273b9b">Lydia Gathoni</a> this session will offer foundastional concepts into machine learning. The session will be <a href="https://meet.jit.si/TheExploreSeriesMachinelearningTrack">available here</a> from 9:30 am.</p><p>Zakia Mustafa is a Data Scientist based in Nairobi, Kenya. She is interested in using Machine Learning and Artificial Intelligence to solve problems. She is also passionate about Technology and how it can be used to transform the African continent. Zakia is currently working on projects in the areas of Computer Vision and Reinforcement Learning, contributing to open source projects and mentoring people who are getting into Data Science. Previously, she worked on machine learning and IOT projects.</p><p>Lydia Gathoni is a Data Scientist and Machine Learning Engineer focused on using data to make predictive models and solve problems. She is a graduate of Kabarak University with a Bsc. in Computer Science majoring in Software Engineering. She has over 4 years of experience in software development where she has worked on several projects and developed systems in various fields and websites. She is passionate about what she does and is driven to deliver innovative solutions that fit the needs of organizations, institutions and society at large.</p><h4>Explore Building Products that Scale.</h4><p>Artificial Intelligence, Machine Learning and Data Science offer us an excellengt opportunity to innovate for the future of work. Join <a href="https://medium.com/u/5407401e5113">Kenneth M Kinyanjui</a> as he facilitates a session on how to build products that scale. The session will begin at 2:30 pm and will be <a href="https://meet.jit.si/TheExploreSeriesProductSchool">available here</a>.</p><p><a href="https://www.linkedin.com/in/kennethkinyanjui/">Kenneth</a> is Software Engineer &amp; Product Manager who can’t get away from solving problems in Emerging Markets. Starting off his career as an independent consultant for companies in Nairobi, San Francisco, London and Seoul he quickly developed into both a Cloud Architect and Infrastructure Engineer. With this new found knowledge and doing this in the luxury of his dorm room, he focused on 2 things. Empowering the community and becoming an entrepreneur. The past 6 years, Kenneth has spent time seeding back to the Sub-Saharan Startup and Technology ecosystem through mentorship, speaking at events , building communities and facilitating workshops. On the other end co-founding 2 companies. WeCashUp (Fintech) based in Marseille, France and Faidika Technologies (Technology Consulting ) based in Kenya.</p><p><strong>PS: </strong>To access the sessions, Kindly click on the respective links on Saturday at the listed times. This <a href="https://forms.gle/cvFTtctCBTs4niWT6">short survey</a> will help us offer support towards this Saturday’s event, please take a minute to fill it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*wUy3i7-odG_djp3kGUmOMw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*7nqA292q4J7wAc0qgzqPbg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*PIeibr9T994jm8aemybI_w.png" /><figcaption>The Explore Series -2020</figcaption></figure><p>One last thing, we are building a community organizers team to help co-ordinate community events, be sure to ping me (muthoniwanyoike@gmail.com) in case you are interested &amp; haven’t done so yet.</p><p>We can’t wait to see you on Saturday!</p><p>Stay safe.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4JpUXCLwWcdfLg8OPAGT7w.jpeg" /><figcaption>The Explore series — 2020</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=daa63af1a98f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/nairobiwimlds/the-explore-series-meet-the-faculty-daa63af1a98f">The Explore Series: Meet the Faculty</a> was originally published in <a href="https://medium.com/nairobiwimlds">NWiMLDS</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[NLP Resources for African Languages — Luganda/Kinyarwanda Translation Model]]></title>
            <link>https://medium.com/nairobiwimlds/nlp-resources-for-african-languages-luganda-kinyarwanda-translation-model-f17b5f62d2a2?source=rss----a090e1dd8a47---4</link>
            <guid isPermaLink="false">https://medium.com/p/f17b5f62d2a2</guid>
            <category><![CDATA[word2vec]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[nlp]]></category>
            <dc:creator><![CDATA[Kathleen S.]]></dc:creator>
            <pubDate>Tue, 26 Feb 2019 04:44:21 GMT</pubDate>
            <atom:updated>2019-02-26T04:44:21.590Z</atom:updated>
            <content:encoded><![CDATA[<h3>NLP Resources for African Languages — Luganda/Kinyarwanda Translation Model</h3><p>About four months ago, I started out on a journey to slowly start building Natural Language Processing tools for African languages. The overall aim is to remove language as a social barrier to success in the global economy of the 21st century. There is a need for people all over the world to be able to use their own language to learn, and especially when using computers or accessing information on the Internet. If you’d like to read my initial thoughts at the start of all this, here’s the <a href="https://www.linkedin.com/pulse/nlp-resources-african-languages-kathleen-siminyu/">previous post</a>.</p><p>I spent the better part of the past four months doing some research. The major language groups on the African continent are as follows(source — <a href="https://en.wikipedia.org/wiki/List_of_ethnic_groups_of_Africa">Wikipedia</a>);</p><ul><li>The Niger-Congo family with a population of 800 million. Major groups include Amhara, Hausa, Oromo, Somali, Tachelhit, Berber and they cover North Africa, the Horn of Africa and the Sahel regions</li><li>The Afro-Asiatic family with a population of 100 million. Major groups include Akan, Fula, Igbo, Kongo, Mande, Mooré, Yoruba, Zulu and Swahili covering the West Africa, Central Africa, Southern Africa and East Africa regions</li><li>The Nilo-Saharan family with a population of 60 million. Major groups include the Dinka, Kanuri, Luo, Maasai and Nuer occupying the Nile Valley, Sahel and East Africa regions</li><li>The Khoisan family with a population of 1 million. Major groups include Nama, San, Sandawe and Kung ǃXóõ located in Southern Africa, particularly Tanzania</li></ul><p>Groups such as the Autronesian and the Indo-European are also notably widely spoken in Africa but due to their origin and prominence elsewhere in the world, I chose not to focus on them. Below is a map with the geographic distributions of these languages.(Source: <a href="https://tracingafricanroots.com/maps/ethno-linguistic/">Tracing African Roots blog</a>)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/553/0*BGMXefHmOcA-HKkx" /></figure><p>Whereas in my previous post, I mentioned the intent to begin with Luhya, Kiswahili and Amharic, initial research made me begin to consider the languages not as independent entities, but rather as members of their larger ethno-linguistic groups where a lot of grammatical characteristics and vocabulary are shared. Leveraging on the similarities of languages in the same linguistic families opens up possibilities for exploring the use of transfer learning to share lexical and sentence level representations across multiple languages further along the line. That being said, I chose the Niger-Congo family, where the Bantu languages fall, for my initial focus, and particularly languages in the J and S zones according to the <a href="https://en.wikipedia.org/wiki/Guthrie_classification_of_Bantu_languages">Guthrie Classification of Bantu Languages</a>. These include Luganda, Chiga, Nyankore, Soga, Haya, Luhya, Nande, Nandi, Kinyarwanda, Rundi, Ha, Shona, Ndau, Sepedi, Sesotho, Tswana, South-Ndebele, North-Ndebele, Xhosa, Zulu, Swazi, Tsonga, Tswa, Ronga, Kalanga, Nambya and Venda. I am sure there are others.</p><p>My first project has been to build a Luganda-Kinyarwanda translation model based on word vector embeddings.</p><p>Word embeddings: Word embedding is the collective name for a set of <a href="https://en.wikipedia.org/wiki/Language_model">language modeling</a> and <a href="https://en.wikipedia.org/wiki/Feature_learning">feature learning</a> techniques in <a href="https://en.wikipedia.org/wiki/Natural_language_processing">natural language processing</a> (NLP) where words or phrases from the vocabulary are mapped to <a href="https://en.wikipedia.org/wiki/Vector_(mathematics)">vectors</a> of <a href="https://en.wikipedia.org/wiki/Real_numbers">real numbers</a>. Conceptually it involves a mathematical <a href="https://en.wikipedia.org/wiki/Embedding">embedding</a> from a space with one dimension per word to a continuous <a href="https://en.wikipedia.org/wiki/Vector_space">vector space</a> with a much lower dimension. (definition from Wikipedia)</p><p>Step 1: Scraping some data from the Internet</p><p>I identified a <a href="https://www.jw.org/en/">website</a> with Bible translations in several languages and was pleasantly surprised to find loads of translations to African languages. Perhaps I should not have been too surprised given that exploration of the African continent was first done by missionaries and in a bid to evangelize, they began learning the local languages, teaching English and translating religious texts. Unfortunately, due to the terms and conditions of use of the site, I am unable to share the the actual text data. Their terms prohibit scraping or harvesting of data for distribution purposes. However, if you are interested, here is a <a href="https://github.com/adtsegaye/crawl-for-parallel-corpora">link</a> to the tool that I used to scrape some data from the site. Feel free to do so for your own personal use.</p><p>As my current interest is in Bantu languages and, for now, particularly those in the J and S zones according to Guthrie’s Classification, I’ve got versions in Luganda, Rundi, Sesotho, Shona, Kinyarwanda, Sepedi, Setswana, Xhosa, Ndebele and IsiNdebele. Plus Kiswahili and English.</p><p>Step 2: Cleaning and tokenizing the data</p><p>I needed to do some basic preprocessing of the text. This involved removing punctuation from the texts, making everything lowercase, removing the numbers that marked chapters and verses of various books of the Bible and finally tokenizing the texts.</p><p>Tokenize: to break text into individual linguistic units</p><p>This process is made fairly simple by a Python library called <a href="https://www.nltk.org/">NLTK</a> for computational linguistics. Here’s the <a href="https://gist.github.com/Witty-Kitty/08d0a79177d004a379a16ea14a05c146">preprocessing code</a> in a github gist.</p><p>Step 3: Training of vector word embedding models</p><p>I came across <a href="https://nlp.stanford.edu/projects/glove/">GloVe</a>, a model for learning word vector representations, and decided to give it a try. I went ahead and trained embeddings for Luganda and Kinyarwanda. There’s several steps to achieving this.</p><p>The <a href="https://github.com/stanfordnlp/GloVe/blob/master/src/vocab_count.c">vocab_count</a> script allows one to extract unigram counts from text data. Here’s the top 20 or so for English and Luganda, just to give you a sense of what the output is like. You get a file with each line containing a word and next to it the number of times that the word has appeared in the text you are working with.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/374/0*CwpV2McjRavlGgFb" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/554/0*QMWrjL6dfsLvMDAw" /></figure><p>Next is finding out the word-word co-occurrence statistics of our corpus using the <a href="https://github.com/stanfordnlp/GloVe/blob/master/src/cooccur.c">cooccur</a>script. This is essentially analyzing the context within which a word is most often used by taking into account it’s neighbouring words. The output of this step is a binary file.</p><p>Here’s a simplistic example to try and explain that concept. From the sentences below, our coocurrence statistics would pick up on the fact that the pronoun ‘I’ is more often than not used alongside a verb as opposed to, say, a noun, because the last sentence is incorrect and should not exist.</p><pre>I am a girl.<br>I saved the world.<br>I cows are always hungry.</pre><p>Excerpt from Glove documentation to further(better) explain co-occurence: <em>As one might expect, ice co-occurs more frequently with solid than it does with gas, whereas steam co-occurs more frequently with gas than it does with solid. Both words co-occur with their shared property water frequently, and both co-occur with the unrelated word fashion infrequently.</em></p><p>I used the default window of 15 words to arrive at the co-occurence statistics for this particular experiment.</p><p>The <a href="https://github.com/stanfordnlp/GloVe/blob/master/src/shuffle.c">shuffle</a> tool is used to shuffle entries in the word-word coocurrence binary files and once again outputs binary files which we can then use to train our GloVe models.</p><p>The <a href="https://github.com/stanfordnlp/GloVe/blob/master/src/glove.c">glove</a> tool takes as input the vocabulary count files that are output of the vocab_count script as well as the shuffled binary files that are the output of the shuffle script and gives, as output, the glove model. I chose to have the model represented as a .txt file with each word and its vector values. Below is sample output.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*PRBL5stpbA9lLyVR" /></figure><p>The first line of each file contains the count of words/tokens whose vector representations the file contains and the number of dimensions of the vectors. In this case, my Luganda file has 42,262 tokens, my Kinyarwanda file has 53,096 tokens and both files contain 50 dimension vectors.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*mBcpihr2jvRITlw8" /></figure><p>Step 4: Aligning the word vector embeddings</p><p>What I achieved in step 3 is to represent the words in my text corpora in a vector space. This however does nothing for my attempts at building a translation model unless these vector embeddings are aligned. If the vocabulary related to ‘family’ is located in the north-west region of my Luganda vector space while in the south-east region of my Kinyarwanda vector space, I need to first align them, have the vector representations overlap, before I can begin the task of translation based on the vectors of each word.</p><p>The family example is simplistic, saying north-west and south-east implies a 2-dimensional space whereas in this instance, my vectors have 50 dimensions. If the latter were the case, aligning the vectors may have been a simple case of plotting and visualizing both then rotating the vectors until they align.</p><p>Given the high dimensions in question, I used <a href="https://github.com/facebookresearch/MUSE">Facebook MUSE</a>’s unsupervised learning approach to learn a mapping between the source space(luganda vectors) and the target space(kinyarwanda). The output is once again .txt files with the various words and their vectors, only this time the vectors for the output file are aligned to those of the source language. As I have no previously built dictionaries for the two languaes, the only means left for me to do some degree of testing of the models is by visualizing the word vectors.</p><p>Step 5: Visualizing the word vector embeddings</p><p>For this step, I used <a href="https://github.com/RaRe-Technologies/gensim">gensim</a> to read in the vector models as KeyedVectors and then proceeded to use a combination of <a href="https://github.com/RaRe-Technologies/gensim">matplotlib</a>, python plotting library, and <a href="https://scikit-learn.org/stable/modules/generated/sklearn.manifold.TSNE.html">t-SNE</a> from sklearn, a tool used to visualize high dimensional data.</p><p>Visualizing all of the output looked something like this for Luganda…</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ouNqIDL743RdrnAZ" /></figure><p>There were just too many tokens, 42,262 to be exact, so I had to purposefully pick out a handful of tokens of interest and then went ahead and visualized those.</p><p>Here’s a visual of some random words I chose to plot. The first, cleaner visualization is simply the words in Kinyarwanda and in Luganda as well. In some cases, I found more than one word for the same thing so the count is not necessarily equal. Luganda words are in green while Kinyarwanda words are in blue.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/942/0*uT-N6DjNdYN2s22n" /></figure><p>The second, slightly busier image is the same visualization except with English translations next to the words. If the word has a ‘null’ next to it, that just means I never got round to getting its translation, the process of doing that was a very tiring and annoying manual process so I did the bare minimum.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/994/0*GLVVP7u9GKB_dVGk" /></figure><p>Finally, here’s the same image another 5 times, each time highlighting a word of interest that appeared in the same ‘neighbourhood’ where vector spaces are concerned and made me very excited at my little win.</p><ul><li>nose(eng), nnyindo(lg), izuru(rw)</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/994/0*mEABrVucC-PiT-CX" /></figure><ul><li>to love each other(eng), kwagalana(lg), gukundana(rw)</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/994/0*Pi4SlgJ72luolqVy" /></figure><ul><li>thing(eng), ikintu/ekintu/engoye(lg), kintu(rw) — besides being in the vicinity of each other in the vector space, the similarity in vocabulary to some extent validates my hypothesis that systematically building resources for closely related languages at a go is a good idea.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/994/0*Mg7Z2RjXmT1ersfy" /></figure><ul><li>one(eng), emu(lg), rimwe(rw)</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/994/0*3bXizLocuSLgbxGa" /></figure><ul><li>name(eng), erinnya(lg), izina(rw)</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/994/0*53klxT3mwfvMF950" /></figure><p>…aannnndd, there you have it folks. Of course there are other instances where the translations are quite unintuitive, let’s attribute that to the fact that I did not work too hard at refining the model. Here’s a gist with the <a href="https://gist.github.com/Witty-Kitty/c82a4213794a4328e765d182c698690e">visualizing code</a>.</p><p>If you’re reading this and wondering how you can help, here’s a couple of ways you could…</p><ul><li>If you know of any websites with text translated to any bunch of African languages, please point me in their direction in the comments section, particularly Bantu languages in the J and S zones, but really just all African languages</li><li>If you are a native speaker of any of the languages I am currently working on and can spare half an hour or so to help with the task of dictionary building, I’m starting to build out small dictionaries which would make future testing/validation and visualization of models easy. Again please reach out in the comments or via a message. I’m currently working through the details of how this will work efficiently</li><li>If you see any other way you can plug in and help out, also ping me</li></ul><p>As next steps, I’m going to be working on tools to better tokenize Bantu languages. Conventional tools are wanting in this instance because Bantu languages tend to be agglutinative. In many instances, the subject, verb, object and various qualifiers can be contained in one word whereas the equivalent in a European language would be a sentence made up of several words. I am also going to work on aligning the texts I already have at sentence level and vary the creation of word vector embeddings to see if I can get better models. Feel free to leave recommendations of any tools, techniques, papers and previous work you think I should check out.</p><p>That’s it for now. Let’s hope it isn’t another 4 months before you hear from me again.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f17b5f62d2a2" width="1" height="1" alt=""><hr><p><a href="https://medium.com/nairobiwimlds/nlp-resources-for-african-languages-luganda-kinyarwanda-translation-model-f17b5f62d2a2">NLP Resources for African Languages — Luganda/Kinyarwanda Translation Model</a> was originally published in <a href="https://medium.com/nairobiwimlds">NWiMLDS</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>