Spotify is currently the worlds leading music streaming ser-vice. As the leader in music streaming the task of providing listeners with music recommendations is vital for Spotify. Listening to playlists is a popular way of consuming music, but traditional recommender systems tend to fo-cus on suggesting songs, albums or artists rather than pro-viding consumers with playlists generated for their needs. This thesis presents a scalable and generalizeable approach to music recommendation that performs song selection for the problem of playlist generation. The approach selects tracks related to a playlist theme by finding the charac-terizing variance for a seed playlist and projects candidate songs into the corresponding subspace. Quantitative re-sults shows that the model outperforms a baseline which is taking the full variance into account. By qualitative results the model is also shown to outperform professionally curated playlists in some cases.

In the age of social networking and music streaming, playlists are a common tool for organizing,sharing or exchanging music in the digital realm. Most research, however, emphasizes mainlypolitical, legal, and ethical constraints of music sharing practices yet, neglects their social impact.Thus, this paper investigates the social-psychological foundation of the playlist and analyses itsfunctionality in establishing social relations and communication. Following the theories of Cooley,Mead, Simmel and Solomon, I conducted and analysed interviews with young Swedish men andwomen, in which they talked about their experiences and attitudes towards playlists. Moreover, allparticipants compiled their own personal playlist, based on certain personality traits, which wereaimed to be recognized by the others during the focus group discussion. The analysis of the datayields the following conclusions: The playlist is a social object, facilitating new forms ofcommunication. The social nature of the playlist is based on the transformation from objective- intosubjective culture. By internalizing new technologies, such as the playlist, objects gain social value,thus mere musical content becomes a social form. It is through sharing and exchanging musicalcompilations that the playlist, as a social form, serves as a vehicle or medium, facilitating newforms of sociation and communication. The communicative function of the playlist is due to itsconstruction through emotions as uniquely subjective judgements, based on the “I” as an emotionalself-feeling. Thus, musical compilations take part in the self-construction process, and can serve asa tool for the symbolic expression of the self.Moreover, the analysis points out that there are differences in how well certain parts of the self canbe communicated by a playlist. Emotional expressions of the self are translated into particularuniversal music patterns most successfully. Furthermore, the analysis shows that some people liketo browse through the playlists of others and judge them thereupon, which results in some type ofmusical voyeurism, termed “playlistism.” In conclusion, I argue that the musical playlist is both,socially implicated and socially implicating, and facilitates communication not only betweenSwedish youth but across cultural boarders.

音樂歌單是由一組多首不同元素、風格的音樂所組成的，它包含了編輯者的個人品味以及因應主題、目的性產生而成。我們可以透過樂曲的律動、節奏、歌曲的主題精神，進而編輯一個相應契合的系列歌曲。當今的音樂收聽市場主要是在網路串流平台上進行隨時、隨地的聆聽，主要的平台有Spotify、Apple Music 以及KKBOX。各家業者不單只是提供使用者歌曲的搜索、單曲的聆聽，更提供訂閱專業歌單編輯者的歌單訂閱服務，甚至是讓一般的使用者參與歌單自訂編輯的過程。然而如何在有限的時間內針對使用者的聆聽習慣去介紹平台上豐富的音樂資源是個很大的挑戰。上述的過程我們稱之為推薦，而當前的音樂推薦研究大多是在對使用者進行相關歌曲的推薦，鮮少能進一步在更抽象層次上的歌單上進行推薦。這邊我們就此一推薦應用提供嵌入式向量表示法學習模型，在有著使用者、歌曲、歌單的異質性社交網路上，對使用者進行歌單的推薦。為了能有效的學習出歌單推薦的模型，我們更將使用者、歌單和歌曲的異質性圖形重組成二分圖（bipartite graph）， 並在此圖形的邊上賦予不等的權重，此一權重是基於使用者隱式反饋獲得的。接著再透過隨機漫步（random walk），根據邊上的權值進行路徑的抽樣選取，最後再將路徑上經過的節點進行嵌入式向量表示法的學習。我們使用歐幾里德距離計算各節點表示法的鄰近關係，再將與使用者較為相關的歌單推薦給使用者。實驗驗證的部分，我們蒐集KKBOX 兩年份的資料進行模型訓練並進行推薦，並將推薦的結果與使用者所喜愛的歌單進行準確度（Precision）評估， 結果證實所得到的推薦效果較一般熱門歌單的推薦來的好，且為更具個人化的歌單推薦。 / Music playlist is crafted with a series of songs, in which the playlist creator has controlled over the vibe, tempo, theme, and all the ebbs and flows that come within the playlist. To provide a personalization service to users and discover suitable playlists among lots of data, we need an effective way to achieve this goal. In this paper, we modify a representation learning method for learning the representation of a playlist of songs, and then use the representation for recommending playlists to users. While there have been some well-known methods that can model the preference between users and songs, little has been done in the literature to recommend music playlists. In light of this, we apply DeepWalk, LINE and HPE to a user-song-playlist network. To better encode the network structure, we separate user, song, and playlist nodes into two different sets, which are grouped by the user and playlist set and song as the other one. In the bipartite graph, the user and playlist node are connected to their joint songs. By adopting random walks on the constructed graph, we can embed users and playlists via the common information between each other. Therefore, users can discover their favorite playlists through the learned representations. After the embedding process, we then use the learned representations to perform playlist recommendation task. Experiments conducted on a real-world dataset showed that these embedding methods have a better performance than the popularity baseline. In addition, the embedding method learns the informative representations and brings out the personal recommendation results.

This thesis analyzes the YouTube service and describes its main deficiencies. Based on theoretical methods and analyses, its main goal is to design a service that will solve the main YouTube problems, build a company around this service and introduce this service to the market. This service will not replace YouTube, but it will supplement it. Further, this work will suggest a possible structure, strategy and information strategy of this new company and its estimated financial results in the first few years.

The master’s thesis deals with testing the application protocol HLS and RTSP in JMeter program. The aim of this thesis is to design and implement a test modules for both protocols, which will perform stress tests. The first part of thesis describes the types of stress tests, JMeter program for performance testing and video on demand services. Next chapter describes selected protokols, especially HLS and RTSP, which are used in this thesis. The practical part contains the design and implementation of test modules including test plans. Finally, the results are processed and commented.

Ce mémoire est composé de trois articles qui s’unissent sous le thème de la recommandation musicale à grande échelle. Nous présentons d’abord une méthode pour effectuer des recommandations musicales en récoltant des étiquettes (tags) décrivant les items et en utilisant cette aura textuelle pour déterminer leur similarité. En plus d’effectuer des recommandations qui sont transparentes et personnalisables, notre méthode, basée sur le contenu, n’est pas victime des problèmes dont souffrent les systèmes de filtrage collaboratif, comme le problème du démarrage à froid (cold start problem). Nous présentons ensuite un algorithme d’apprentissage automatique qui applique des étiquettes à des chansons à partir d’attributs extraits de leur fichier audio. L’ensemble de données que nous utilisons est construit à partir d’une très grande quantité de données sociales provenant du site Last.fm. Nous présentons finalement un algorithme de génération automatique de liste d’écoute personnalisable qui apprend un espace de similarité musical à partir d’attributs audio extraits de chansons jouées dans des listes d’écoute de stations de radio commerciale. En plus d’utiliser cet espace de similarité, notre système prend aussi en compte un nuage d’étiquettes que l’utilisateur est en mesure de manipuler, ce qui lui permet de décrire de manière abstraite la sorte de musique qu’il désire écouter. / This thesis is composed of three papers which unite under the general theme of large-scale music recommendation. The first paper presents a recommendation technique that works by collecting text descriptions of items and using this textual aura to compute the similarity between them using techniques drawn from information retrieval. We show how this representation can be used to explain the similarities between items using terms from the textual aura and further how it can be used to steer the recommender. Because our system is content-based, it is not victim of the usual problems associated with collaborative filtering recommenders like the cold start problem. The second paper presents a machine learning model which automatically applies tags to music. The model uses features extracted from the audio files and was trained on a very large data set constructed with social data from the online community Last.fm. The third paper presents an approach to generating steerable playlists. We first demonstrate a method for learning song transition probabilities from audio features extracted from songs played in professional radio station playlists. We then show that by using this learnt similarity function as a prior, we are able to generate steerable playlists by choosing the next song to play not simply based on that prior, but on a tag cloud that the user is able to manipulate to express the high-level characteristics of the music he wishes to listen to.

台灣廣播媒體生態自進入頻道開放以來，目前除電台數目增加為118家以外，也出現了許多以音樂頻道為經營型態的類型音樂電台。過去部分的音樂文化論述在談及廣播電視媒體時，論點不乏多所批評其與唱片公司勾結來配合打歌宣傳，對於音樂文化的負面影響甚巨。但是如今，電台經營越趨專業化的情況下，似乎在台灣，廣播電台與流行音樂工業間的關係，也逐漸有了轉變。 本研究，藉由音樂社會學以及媒介產製兩種研究取向，各取其關切的焦點。來探究當今台灣出現越來越多的音樂電台，在當代的音樂文化生產體制上，佔據了何種角色？地位？以及影響層面？並且藉由區分不同的音樂電台間在制度上所造成的結果不同，作為重新認識廣播媒體與音樂文化之間互動關係的推論知識。 在當今，幾乎所有的電台都有一套選擇音樂的音樂決策，有些電台也建立了播出單制度，音樂能否進入播出單？在電台方面，組織因素取決於電台自身再經營類型上的選擇，其中也包括經理人、音樂總監、節目主持人在工作權限上相互影響。外在影響上，則受到唱片公司唱片促銷的影響，多數唱片公司與電台在音樂輸出方面，兩者間存在著複雜的互動、包含了廣告運作、唱片促銷人員與電台工作人員間的人際因素。此外，台灣當代音樂文化的變遷，所造成的當今音樂呈現了哪些文化上的特質？各種音樂受喜愛的程度？也影響了音樂是否能夠被電台選擇進入播送？ 經由上述的各不同面向，屬於台灣音樂電台的的文化生產機制是如何運作？能否按照以往音樂文化論述者的觀點？研究結果顯示，目前的音樂電台，已經走向專業化經營、對於音樂多有一套管理、選擇的機制運作，但多數流行音樂電台與唱片業界仍具有一種合則兩利的共謀關係，藉由與唱片業的廣告合作，將唱片促銷置入了電台的播歌高度循環率中。另一方面，仍有其他電台，並不需要倚賴唱片公司，在節目呈現與經營上皆有其獨特性。因此，區分不同電台間的差別格外重要。 結論認為，台灣目前出現的音樂電台類型，是以音樂市場性為基礎，但面臨唱片公司、電台組織、音樂文化變遷三者的交互影響下，電台在文化生產的角色中，不該被認為是單純的通道，而是具有積極的影響力，以往對於廣播媒體在音樂生產方面的批評焦點，應轉至電台是否能保持傳播媒體的傳播性，而非過度跟隨主流市場。

Ce mémoire est composé de trois articles qui s’unissent sous le thème de la recommandation musicale à grande échelle. Nous présentons d’abord une méthode pour effectuer des recommandations musicales en récoltant des étiquettes (tags) décrivant les items et en utilisant cette aura textuelle pour déterminer leur similarité. En plus d’effectuer des recommandations qui sont transparentes et personnalisables, notre méthode, basée sur le contenu, n’est pas victime des problèmes dont souffrent les systèmes de filtrage collaboratif, comme le problème du démarrage à froid (cold start problem). Nous présentons ensuite un algorithme d’apprentissage automatique qui applique des étiquettes à des chansons à partir d’attributs extraits de leur fichier audio. L’ensemble de données que nous utilisons est construit à partir d’une très grande quantité de données sociales provenant du site Last.fm. Nous présentons finalement un algorithme de génération automatique de liste d’écoute personnalisable qui apprend un espace de similarité musical à partir d’attributs audio extraits de chansons jouées dans des listes d’écoute de stations de radio commerciale. En plus d’utiliser cet espace de similarité, notre système prend aussi en compte un nuage d’étiquettes que l’utilisateur est en mesure de manipuler, ce qui lui permet de décrire de manière abstraite la sorte de musique qu’il désire écouter. / This thesis is composed of three papers which unite under the general theme of large-scale music recommendation. The first paper presents a recommendation technique that works by collecting text descriptions of items and using this textual aura to compute the similarity between them using techniques drawn from information retrieval. We show how this representation can be used to explain the similarities between items using terms from the textual aura and further how it can be used to steer the recommender. Because our system is content-based, it is not victim of the usual problems associated with collaborative filtering recommenders like the cold start problem. The second paper presents a machine learning model which automatically applies tags to music. The model uses features extracted from the audio files and was trained on a very large data set constructed with social data from the online community Last.fm. The third paper presents an approach to generating steerable playlists. We first demonstrate a method for learning song transition probabilities from audio features extracted from songs played in professional radio station playlists. We then show that by using this learnt similarity function as a prior, we are able to generate steerable playlists by choosing the next song to play not simply based on that prior, but on a tag cloud that the user is able to manipulate to express the high-level characteristics of the music he wishes to listen to.