白拾的笔记本

Notice Use CUE as baseline. Ideas Use information theory to explain the process. Plan [ ] Use uncertainty estimation method in CUE. [ ] Evaluate CUE on semantic kitti (outdoor), or scannet v2 (indoor). [ ] Extract uncertainty estimation method of CUE. [ ] Achieve socket communication for model inference for CUE model. [ ] Socket communication. [ ] Local small model training, calculate uncertainty information. [ ] Cloud large model training, with global feature map and compressed point c ...

Helpful Materials: GitHub repo https://github.com/ramdrop/cue https://github.com/NVIDIA/MinkowskiEngine

65f8e3b87362ab224b1b50bf77ce06905769e529ac2ce2fb173717936f4bfa951388d187066e6a12bd1c6d157f10db3ac08f94aec10dc56957cf4568bcd5edf1fe18ac358c01221957ce93e89a26a747ba1af9802df952fe1f1d543d22a7cfb7bd0d82ddd31b20e1d206d81a12407762ee3bb3ae12061442e5e395a852bd19418635a85aad4bea1f660d0aa75db3aa1fd6f66c102ce7c45c13e9946238414bd85f9074c88e756e207414de2f1a47e80a9cd5a3dba30b227dd52e61cc9e8d2d40911f1512caddd8761d482c308677ee545682561ddbc96a2e9e34920b1a28d69fce3f218b3cab57306771d3deb4289bfc1a51e37d40eef70d0 ...

65f8e3b87362ab224b1b50bf77ce06905769e529ac2ce2fb173717936f4bfa95dee07c63ee55b1eb7d727a3289677f4e37dcace7e04c16f7baac110a3f6bf5066bec759045f493fc06132199b6550ed963c6b2ac1c832d53a7125caf8084d2d075494af4d3c2334f62544b8bb009f357d8f653605eb98746bf5b7aa0b092ee0179b908dd9797c9c5ad85bd2544d5424e42519ac5df7bc43944d2c6f9ad35c9ccbd9ee4b93da63d2a6ee64422f9c36a97297e19e2b0006b11f6602d73a92600d0c372137798094280e657c99be97a14b26319a447a50ce9660c36ec697c56c2b18382f32ab20e860fccdc5b62681d228b5c994f0e6486f0880 ...

65f8e3b87362ab224b1b50bf77ce06905769e529ac2ce2fb173717936f4bfa953ddf2337fef7e8935e7f6faf9d9e071ac3cb5cfd8a6a29534f7778efa6a511c8fccceb9b420be6b9286cc5b475cc232af853d6664ecd2df92e962fcfdf962a401c1106344ab8dbbc60a31840ee07c5cdc04e25cc8569c95df1d1dde0eb0d3761c5ba24d853cc2e8bb19ef33c3115a967a287afd9d973d53c4f66db0050432713722529bbf179c9b177a904f2ace9bccf7a3005de18a4707f8b7630412b5d401ecef5caf4907012584a84d8be70e86806b832acdf345719df4264954bbb97cf112f124b71bb48b3544e61092bddf273630917fefc975c9a181 ...

From subword units to n-grams: hierarchy of models Defining a hierarchy of models: we can compile different HMMs to create models of utterances We can do some pruning, remove some tokens while proceeding, reduce computation cost (Maybe Heuristic is also can be helpful in such case.) Conditional independence and the forward algorithm We use the Markov property of HMMs (i.e. conditional independence assumptions) to make computing probabilities of observation sequences easier HMM tr ...

Hidden Markov Models for ASR Intro to Hidden Markov Models, comparison to DTW. dynamic time warping (DTW) Hidden state sequences and alignment Hidden state sequence, Trellis (i.e. lattice) representation of an HMM, aligning observations to states grid and lattice The Viterbi algorithm and token passing The Viterbi algorithm can be computed on different data structures. The token passing version turns out to be very convenient for ASR. Tthese two different implementations are simply two differe ...

Gaussian distributions in models Classification model If features are highly correlate with each other, we can solve this correlation by rotating the axis, by PCA. Gaussian distribution of classification result of feature vector Decision boundary Cepstral Analysis, Mel-Filterbanks We now start thinking about what a good representation of the acoustic signal should be, motivating the use Mel-Frequency Cepstral Coefficients (MFCCs). Since the feature in a feature vector is correlated, if we want ...

Mission, scenarios and behaviors Task: Navigate the robot itself from the current position to final destination on the map, without human intervention. Goal: find most efficient path (trajectory) in terms of time or distance travelled. Example Common Behavior Sets Speed tracking Deceleration to stop Stay stopped Yield Emergency stop Motion planning constraints Read more about vehicle dynamic. Objective functions in planning Supplementary Readings Motion Planning Fo ...

Cochlea, Mel-Scale, Filterbanks From human speech perception to considerations for features for automatic speech recognition Cochlea Different places along the cochlea respond to the incoming frequency. Mel scale Nonlinearity in hertz scale, linear in mal scale, for cochlea. Filter banks Simplify of cochlea is like a bank of bandpass filters. lower frequency limit and higher frequency limit. Wider and wider in higher frequency. triangular filters is more appropriate than the rectangle on ...

Connected and Citation Speech Connected speech differs from the citation form. Connected Speech Processes Connected speech forms are highly variable as the result of a number of processes that apply to consonants and vowels. Prosodic Structure Prosody is the combination of speech properties that break speech into units of time, indicate the boundaries of those units, and highlight certain constituents. A constituent is a word or a group of words that function as a single u ...

Module 5 TTS front-end We want to generate speech that is Intelligible: you can clearly perceive what words are being said Natural: sounds like human speech Appropriate: conveys the right meaning in a specific context Front-end: Analyze text, generate a linguistic specification of what to actually generate Front-end purpose: derive a linguistic specification from text that includes the necessary information to generate speech Back-end: Waveform generation from the linguistic specificatio ...

Occupancy grid Occupancy Map Calculus Practicle issues of Occupancy map Inverse Measurement Model Downsampling for lidar Other Type of Map Supplementary Readings Note of Occupancy Maps. (MUST READ) Lanelets: Efficient Map Representation for Autonomous Driving. (Optional) Probabilistic robotics. Read Chapter 9 - Occupancy Grid Mapping for an overview of how occupancy grids are generated Origin: Dr. Chris Lu (Homepage) Translate + Edit: YangSier (Homepage) :four_leaf_clov ...

LiDAR LiDAR calculus State Estimation via Point Set Regression Problem define ICP Algorithm Origin: Dr. Chris Lu (Homepage) Translate + Edit: YangSier (Homepage) :four_leaf_clover:碎碎念:four_leaf_clover: Hello米娜桑，这里是英国留学中的杨丝儿。我的博客的关键词集中在编程、算法、机器人、人工智能、数学等等，点个关注吧，持续高质量输出中。 :cherry_blossom:唠嗑QQ群：兔叽的魔术工房 (942848525) :star:B站账号：白拾Official（活跃于知识区和动画区）