Is it Normal for a Neural Network Loss to Increase after being trained on an example?





.everyoneloves__top-leaderboard:empty,.everyoneloves__mid-leaderboard:empty,.everyoneloves__bot-mid-leaderboard:empty{ height:90px;width:728px;box-sizing:border-box;
}







0















I am currently testing an LSTM network. I print the loss of its prediction on a training example before back-propagation and after back-propagation. It would make sense that the after loss should always be less than the before loss because the network was just trained on that example.



However, I am noticing that around the 100th training example, the network begins to give a more inaccurate prediction after back-propagation than before back-propagating on a training example.



Is a network expected to always have the before loss be higher than the after loss? If so, are there any reasons this happens?



To be clear, for the first hundred examples, the network seems to be training correctly and doing just fine.






python machine-learning neural-network lstm recurrent-neural-network







share|improve this question























  • Did you try decreasing your learning rate?

    – cheersmate
    Nov 26 '18 at 8:23


















0















I am currently testing an LSTM network. I print the loss of its prediction on a training example before back-propagation and after back-propagation. It would make sense that the after loss should always be less than the before loss because the network was just trained on that example.



However, I am noticing that around the 100th training example, the network begins to give a more inaccurate prediction after back-propagation than before back-propagating on a training example.



Is a network expected to always have the before loss be higher than the after loss? If so, are there any reasons this happens?



To be clear, for the first hundred examples, the network seems to be training correctly and doing just fine.






python machine-learning neural-network lstm recurrent-neural-network







share|improve this question























  • Did you try decreasing your learning rate?

    – cheersmate
    Nov 26 '18 at 8:23














0












0








0








I am currently testing an LSTM network. I print the loss of its prediction on a training example before back-propagation and after back-propagation. It would make sense that the after loss should always be less than the before loss because the network was just trained on that example.



However, I am noticing that around the 100th training example, the network begins to give a more inaccurate prediction after back-propagation than before back-propagating on a training example.



Is a network expected to always have the before loss be higher than the after loss? If so, are there any reasons this happens?



To be clear, for the first hundred examples, the network seems to be training correctly and doing just fine.






python machine-learning neural-network lstm recurrent-neural-network







share|improve this question














I am currently testing an LSTM network. I print the loss of its prediction on a training example before back-propagation and after back-propagation. It would make sense that the after loss should always be less than the before loss because the network was just trained on that example.



However, I am noticing that around the 100th training example, the network begins to give a more inaccurate prediction after back-propagation than before back-propagating on a training example.



Is a network expected to always have the before loss be higher than the after loss? If so, are there any reasons this happens?



To be clear, for the first hundred examples, the network seems to be training correctly and doing just fine.






python machine-learning neural-network lstm recurrent-neural-network




python machine-learning neural-network lstm recurrent-neural-network






share|improve this question













share|improve this question











share|improve this question




share|improve this question










asked Nov 25 '18 at 5:33









Rehaan AhmadRehaan Ahmad

452317




452317













  • Did you try decreasing your learning rate?

    – cheersmate
    Nov 26 '18 at 8:23



















  • Did you try decreasing your learning rate?

    – cheersmate
    Nov 26 '18 at 8:23

















Did you try decreasing your learning rate?

– cheersmate
Nov 26 '18 at 8:23





Did you try decreasing your learning rate?

– cheersmate
Nov 26 '18 at 8:23












1 Answer
1






active

oldest

votes


















0














Is your dataset shuffled?
Otherwise it could be the case that it was predicting one class for the first 99 examples.
If not then LSTM can be tricky to train. Try changing hyper parameters and also I would recommend starting with SimpleRNN, GRU and then LSTM as sometimes a simple network might just do the trick.






share|improve this answer
























    Your Answer






    StackExchange.ifUsing("editor", function () {
    StackExchange.using("externalEditor", function () {
    StackExchange.using("snippets", function () {
    StackExchange.snippets.init();
    });
    });
    }, "code-snippets");

    StackExchange.ready(function() {
    var channelOptions = {
    tags: "".split(" "),
    id: "1"
    };
    initTagRenderer("".split(" "), "".split(" "), channelOptions);

    StackExchange.using("externalEditor", function() {
    // Have to fire editor after snippets, if snippets enabled
    if (StackExchange.settings.snippets.snippetsEnabled) {
    StackExchange.using("snippets", function() {
    createEditor();
    });
    }
    else {
    createEditor();
    }
    });

    function createEditor() {
    StackExchange.prepareEditor({
    heartbeatType: 'answer',
    autoActivateHeartbeat: false,
    convertImagesToLinks: true,
    noModals: true,
    showLowRepImageUploadWarning: true,
    reputationToPostImages: 10,
    bindNavPrevention: true,
    postfix: "",
    imageUploader: {
    brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
    contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
    allowUrls: true
    },
    onDemand: true,
    discardSelector: ".discard-answer"
    ,immediatelyShowMarkdownHelp:true
    });


    }
    });














    draft saved

    draft discarded


















    StackExchange.ready(
    function () {
    StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53464933%2fis-it-normal-for-a-neural-network-loss-to-increase-after-being-trained-on-an-exa%23new-answer', 'question_page');
    }
    );

    Post as a guest















    Required, but never shown

























    1 Answer
    1






    active

    oldest

    votes








    1 Answer
    1






    active

    oldest

    votes









    active

    oldest

    votes






    active

    oldest

    votes









    0














    Is your dataset shuffled?
    Otherwise it could be the case that it was predicting one class for the first 99 examples.
    If not then LSTM can be tricky to train. Try changing hyper parameters and also I would recommend starting with SimpleRNN, GRU and then LSTM as sometimes a simple network might just do the trick.






    share|improve this answer




























      0














      Is your dataset shuffled?
      Otherwise it could be the case that it was predicting one class for the first 99 examples.
      If not then LSTM can be tricky to train. Try changing hyper parameters and also I would recommend starting with SimpleRNN, GRU and then LSTM as sometimes a simple network might just do the trick.






      share|improve this answer


























        0












        0








        0







        Is your dataset shuffled?
        Otherwise it could be the case that it was predicting one class for the first 99 examples.
        If not then LSTM can be tricky to train. Try changing hyper parameters and also I would recommend starting with SimpleRNN, GRU and then LSTM as sometimes a simple network might just do the trick.






        share|improve this answer













        Is your dataset shuffled?
        Otherwise it could be the case that it was predicting one class for the first 99 examples.
        If not then LSTM can be tricky to train. Try changing hyper parameters and also I would recommend starting with SimpleRNN, GRU and then LSTM as sometimes a simple network might just do the trick.







        share|improve this answer












        share|improve this answer



        share|improve this answer










        answered Nov 25 '18 at 8:39









        VitrioilVitrioil

        8915




        8915
































            draft saved

            draft discarded




















































            Thanks for contributing an answer to Stack Overflow!


            • Please be sure to answer the question. Provide details and share your research!

            But avoid …



            • Asking for help, clarification, or responding to other answers.

            • Making statements based on opinion; back them up with references or personal experience.


            To learn more, see our tips on writing great answers.




            draft saved


            draft discarded














            StackExchange.ready(
            function () {
            StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53464933%2fis-it-normal-for-a-neural-network-loss-to-increase-after-being-trained-on-an-exa%23new-answer', 'question_page');
            }
            );

            Post as a guest















            Required, but never shown





















































            Required, but never shown














            Required, but never shown












            Required, but never shown







            Required, but never shown

































            Required, but never shown














            Required, but never shown












            Required, but never shown







            Required, but never shown







            -

            這個網誌中的熱門文章

            Tangent Lines Diagram Along Smooth Curve

            圖片
            7 1 I am trying to replicate a diagram and have a minimal example. I don't know how to add additional tangents between the specified points and preserve smoothness. documentclass{article} usepackage{tikz} usetikzlibrary{calc,intersections} begin{document} begin{tikzpicture} % Axes draw [->, name path=x] (-1,0) -- (11,0) node [right] {$x$}; draw [->] (0,-1) -- (0,6) node [above] {$y$}; % Origin node at (0,0) [below left] {$0$}; % Points coordinate (start) at (1,-0.8); coordinate (c1) at (3,3); coordinate (c2) at (5.5,1.5); coordinate (c3) at (8,4); coordinate (end) at (10.5,-0.8); % show the points foreach n in {start,c1,c2,c3,end} fill [black] (n) circle (2pt) node [below] {}; % join the coordinates draw [thick,name path=curve] (start) to[out=70,in=180] (c1) to[out=0,in=180]
            閱讀完整內容

            Yusuf al-Mu'taman ibn Hud

            圖片
            Main article: Taifa of Zaragoza Al-Mu'taman Ruler of Zaragoza Reign 1081–1085 Predecessor Ahmad al-Muqtadir Successor Al-Mustain II Born Zaragoza Died 1085 Full name Yusuf ibn Ahmad al-Mu'taman ibn Hūd House Banu Hud Yusuf ibn Ahmad al-Mu'taman ibn Hūd (Arabic: المؤتمن بالله يوسف إبن أحمد إبن هود ‎, al-Mutaman bi l-Lah , died c. 1085) was the third king of the Banu Hud dynasty who ruled over the Taifa of Zaragoza from 1081 to 1085. Yusuf al-Mutaman reigned during the height of power of Muslim Zaragoza , following the thriving period of his father Ahmad al-Muqtadir. He continued his father's efforts and created around him a court of intellectuals, living in the beautiful palace of Aljafería, nicknamed as "the palace of joy". Al-Mu'taman was also a scholarly king and a patron of science, philosophy and arts. As an example of a wise king, he knew astrology, philosophy, and especially mathematics, a discipline in which
            閱讀完整內容

            Zucchini

            圖片
            This article is about the vegetable. For other uses, see Zucchini (disambiguation). Zucchini (courgette) A striped and a uniform color zucchini Genus Cucurbita Species Cucurbita pepo Origin 19th century northern Italy The zucchini ( / z uː ˈ k iː n i / , American English) or courgette ( / k ʊər ˈ ʒ ɛ t / , British English) is a summer squash which can reach nearly 1 metre (100 cm; 39 in) in length, but is usually harvested when still immature at about 15 to 25 cm (6 to 10 in). [1] A zucchini is a thin-skinned cultivar of what in Britain and Ireland is referred to as a marrow. [2] [3] In South Africa, a zucchini is known as a baby marrow. Along with certain other squashes and pumpkins, the zucchini belongs to the species Cucurbita pepo . It can be dark or light green. A related hybrid, the golden zucchini, is a deep yellow or orange color. [4] In a culinary context, the zucchini is treated as a vegetable; it is usually cooked and presented as a savory dish o
            閱讀完整內容